Determinants of loan repayment performance of household and micro borrowers in vietnam

It’s very difficult for them to determine which are key factors affecting on loan repayment performance of micro-businesses in Vietnam, especially when using cash in payment is still pop

VIETNAM NATIONAL UNIVERSITY, HANOI

VIETNAM JAPAN UNIVERSITY

-PHAM THANH DUNG

DETERMINANTS OF LOAN REPAYMENT PERFORMANCE OF MICROBUSINESS

THESIS ACKNOWLEDGMENT

Foremost, I would like to express my sincere gratitude to my advisors Dr VuAnh Dung and Dr Hiroshi Morita for the support of my Thesis study andresearch, for their motivation, and immense knowledge, as well their experience

in the field Their guidance has helped me much in the time of research andwriting of this thesis

Besides my advisors, I would like to sincerely thank the rest of my thesiscommittee: Prof and Dr Matsui, Dr Tran Thi Lien, Dr Kodo, Dr Tran Thi BichHanh, and especially Dr Yoshifumi Hino, for their encouragement, insightfulcomments, and hard questions

My thanks also go to Ms Nguyen Thi Huong, MBA program assistant who hasenthusiastically supported me in the process of completing the procedure, as well

as connecting for smooth communication between students and advisors in thetime of my doing research

TABLE OF THE CONTENTS

ABSTRACT

LIST OF TABLES

LIST OF FIGURES

LIST OF DEFINITIONS AND ABBREVIATIONS

CHAPTER ONE: INTRODUCTION 1

1.1 Background of the Problem 1

1.2 Statement of the Problems 4

1.3 Objectives of the Study 5

1.4 Research Questions 6

1.5 Scope of the Study 6

1.6 Structure of the research 6

CHAPTER TWO: LITERATURE REVIEW 8

2.1 Concept and Definition 8

2.2 Related Literature Reviews to the Variables Used in the Study 15

2.3 Cox Regression Model 24

Table 2.1 Comparison of models on the random sample 25

2.4 Research Gap 26

CHAPTER THREE: DATA AND METHODOLOGY

3.1 Description of the Study Area

3.2 Data Description

3.2.1 Sampling procedure and technique

3.3 Variables in the Research 3.3.1 Dependent Variables

3.4 Method for Data Analysis

3.5 Research Model

3.6 Summary of Cox Model factors in this research

CHAPTER 4: RESULTS AND DISCUSSIONS 39

4.1 Introduction 39

4.2 Summary Statistics 39

4.3 Results of the Cox Proportional Hazard Model 41

4.4 Discussions 51

CHAPTER 5: CONCLUSIONS AND RECOMMENDATIONS 61

5.1 Conclusions 61

5.2 Recommendations 63

5.3 Recommendations for Further Research 68

5.4 Limitations of the Study 68

REFERENCES 70

APPENDIX 73

Introduction: Vietnam is a developing economy where the service industry and light

industry account for a big percentage of businesses in the economy Businesses in these sectors are almost small and micro enterprises and household/family businesses (MSEs) They are always in need of capital for their activities However, it’s very difficult for them to get access to loans from banks and credit organizations Because they almost don’t have collateral to secure for their loan and are not able to provide an eligible financial statement which is usually required

by banks Consequently, banks or creditors are not interested in offering loan to such clients because they usually don’t have collaterals and have difficulties in providing documents as required by banks, which is very risky to a bank if they accept such types of customers, whereas, there are not enough tools to define and mitigate relevant risk It’s very difficult for them to determine which are key factors affecting on loan repayment performance of micro-businesses in Vietnam, especially when using cash in payment is still popular, which makes banks find it more difficult and not interested in loans for micro business So, detailed research on the determinants of loan repayment performance of micro-businesses in Vietnam is essential

Objectives: The objectives of this research is to analyze the determinants of credit

risk of loan repayment performance of micro-businesses in Vietnam

Methodology: A study was conducted on 200 enterprises and household businesses

who have taken loans from Banks and P2P lending companies in Vietnam in 24 months starting from Jan 2018 to the end of 2019 (Regarding the average term of a loan is 12 months).

The data used in this study was the secondary data from a joint-stock commercial bank and a P2P Lending company in Vietnam The Cox regression model is used with twelve explanatory variables.

Loan repayment status/Default rate is the dependent variable, while twelve characteristics of borrower and the enterprise owned by the borrower’s characteristics are considered as explanatory variables In this case, the value of the dependent variable (loan repayment status) is 0 and 1, if borrowers defaulted it takes 1 and otherwise 0.

Results: Ten out of eleven significant factors identified through the relevant model

are: Gender, Age, Housing, Educational level, Business sector, Years in business, Percentage of share, Digital sales channel, Number of Clients, and Turnover stability in 6 latest months, are the key factors that affect the loan repayment performance Among them, 5 new factors which are analyzed regarding Vietnam

economy’s characteristics and as a result of this Research are: Percentage of share owning, Digital sales channel, Number of clients, Turnover stability in 6 latest months and Years in Business.

Conclusion: Hence, regarding the Research’s result, Banks and financial

companies can use it to build their credit scoring models or system Accordingly, increase their ability in risk management as well as enhance their desire in helping MSEs get the loan that suitable and necessary for them, which is very important to improve economic growth Policymakers should pay attention to issue more proper policy to further support for MSEs who have been given with not enough financial

as well as non-financial aids Furthermore, enterprise’s stakeholders can use this research’s result to increase performance rating for themselves and improve its rating to banks and creditors.

Keywords: Credit risk; Loan repayment performance; Microcredit; Micro business;

Cox Regression model.

LIST OF TABLES

Table 2.1 Comparison of models on the random sample 25

Table 3.1 Sample Distribution 32

Table 3.2: Description of independent variables 33

Table 4.1: Status of repayment 39

Table 4.3: Summary statistics for continuous variables 41

Table 4.4: Univariate analysis result for each covariate 42

Table 4.5: Multivariate Cox Proportional Hazards Regression results 45

Table 4.6: Final model of Cox PH 46

Table 4.7: Cox Model with Time-Dependent Covariates 47

Table 4.8 Analysis of The variable of “Business sector” as categorical variable 48

LIST OF FIGURES

Figure 1.1 Enterprise size categorizes by capital scale and labor scale 2

LIST OF DEFINITIONS AND ABBREVIATIONS

CHAPTER ONE: INTRODUCTION

1.1 Background of the Problem

Micro enterprises and trading households (MSEs) play a vital role in the economicgrowth of Vietnam MSEs not only contribute to GDP of Vietnam, but also helpcreate millions of jobs which make direct contribution to reduce unemployment andenhance public welfare Vietnam is home to 700,000 enterprises registered under theLaw on Enterprises and more than 5.2 million trading households, in addition toforeign-invested enterprises In this case, the private sector accounts for over 60%

of the GDP, which undoubtedly makes it a driving force of the economy (TheVietnam Investment Review, 2019)

In terms of labor scale, nearly 98% of total enterprises in Vietnam are classified asMicro, Small and Medium Enterprises (MSMEs) while large-sized enterprisesaccount for only the remaining 2% (Decree No 56/2009/ND-CP) Moreover, by theGeneral Statistics Office of Vietnam, in 2017, Vietnam also has more than 5.2million trading households (HHs) which have similar characteristics as microenterprises (GSO, 2016) This research, research subjects are included of bothmicro-enterprises and trading households, which is abbreviated as MSEs in thisdocument

Figure 1.1 Enterprise size categorizes by capital scale and labor scale

Since 2000, when the Enterprise Law was promulgated, Vietnam is emerging as anattractive and potential address for investors from all over the world According tothe Annual Business Report 2016 by Vietnam Chamber of Commerce and Industry,Vietnam witnessed a considerable increase in the number of new enterprisesregistered, number of employees, as well as total capital during the 2007-2015period In line with the remarkable economic development of Vietnam overall, smalland medium-sized enterprises (SMEs) in Vietnam also have experienced amazinggrowth, especially

However, 70-80% MSEs in Vietnam have been facing with many challenges anddifficulties One of key challenges to them is not having access to finance throughformal channels (Banks, other financial companies) According to the Insight Asiasurvey of SMEs in 4 areas of F&B, supporting industry, fashion apparel andcosmetics which have fewer than 200 employees and now have stable productionwith workshops but intend to expand, or relocate its production / service area, 62%

of the respondents said that they faced difficulties in capital (mainly to invest infactories, machinery )

In recent years, it is easier for MSEs to get a loan due to an increase in creditchannels to them Some banks and financial companies such as VPBank, FEcredit,MSB, etc… have started offering loans to MSEs, but the number is still limited.However, upward ratio of bad debt and Non-performing-loan (NPL) has been a bigconcern to many banks and creditors and make them hesitate in making new loans.NPL is a situation in which, all or a part of principal and interest is not paid duly byborrowers which means bank will lose their money.

Consequently, banks will tighten their credit policy and the access to finance for MSEs

is even smaller They require MSEs to give collateral or strict requirement andprocedure or documents to get a loan In the reverse side, MSEs do not have collaterals

or are not in business long enough (usually more than 2 years as required byVietnamese banks) to be provided with unsecured loans, incapable of formulatingfeasible production, business, projects, not good financial recording, etc forcommercial banks to consider loan Therefore, It’s very difficult for MSEs to get loansfrom banks and credit organizations The priority almost is given to state ownedenterprises, or big enterprises, or enterprises which have already doing business dozenyears Due to lack of capital, MSME can’t develop as they should have been, even theyface the risk of bankruptcy, being merged and acquired quite easily

Difficulties piled up, MSEs have no way than "take risk" to find informal sources ofcapital, also known as "black market" with extremely high lending interest rates,which almost lead to undesirable consequences as we already knew through manyreport/articles of violence relating to loans from black market recently As unofficialsurvey by Viet Nam Business Association, nearly 70% of 2,600 surveyed small-and-medium sized enterprises (SMEs) had to seek loans on the black market as theywere unable to access bank loans The businesses’ ability to access loans fell from

45 per cent in 2011-2013 to 24 per cent in 2015 The loans granted to the remaining

30 per cent accounted for only 3 per cent of total bank capital in the market

1.2 Statement of the Problems

The root cause of NPL is internal factors of the MSEs debtors Banks can make amistake in making a loan decision for the debtor because the banks have a difficultyto

separate good quality and the poor quality debtor (Taswan, 2011), which is oftencalled asymmetric information experienced by financial institutions (banks) anddebtor as borrowers (Bakhtiar and Sugema, 2012) and can lead to adverse selection.Wrong choice and disbursement of borrowers will put banks into a double-risksituation: make disbursement to not qualified borrowers and reject qualifiedapplicants To minimize the risk of NPL, by [4], banks in the world implement acreditworthiness test system called Credit Scoring However, the more importantthing is to predict the time-to-default has not yet been well defined In Vietnam, justfew banks focus on credit to MSEs but they almost use traditional ways of case – by– case appraisal They also haven’t implemented a credit Scoring systemsophisticated enough due to inadequate analysis of key determinants and lack of aproper credit risk modelling Whereas, having a credit risk modelling for internalrating is more and more important, especially nowadays, managent of risk isrequired by the standard of Basel II in many countries in the world, by [27]

This leads to a fact that it’s even more difficulty for MSEs to get loans Being afraid

of making wrong loan decisions, banks have trend to increase the requirement orincrease interest rate to compensate for the risk cost that rise from unqualifiedcustomers and risk they may have to incur loss Accordingly, many MSEs find iteven more difficult to get a loan from banks If they can get a loan, the interest rate

is also very high compared with the return that they can make from their business.This, finally, make the flow of capital not smooth and limit economic growth Also,high NPL ratio can also lead to crisis and make banks stop lending out Also, it alsocan make investors worried of putting money in new enterprises as well as depositmoney into bank Accordingly, this affect to all businesses in all fields, hence to thewhole economy

Internationally, there are many research using different models on determinants ofcredit risk of MSEs, but no research is made on trading households which is one ofimportant business force in Vietnam In Vietnam, no research on determinants ofcredit risk of MSEs is available up to now Hence, the research is based on someresearch from foreign countries with similar characteristics in the economy to findout determinants on credit risk of loan repayment performance of micro business inVietnam.

Due to this reason, this Research is necessary to analysis a credit risk of loanrepayment of micro business in Vietnam So that, it helps banks and creditorganizations easier to define key determinants influencing default rate to mitigaterisk for themselves and widely open to offering loan to MSEs in the economy.Accordingly, it will help banks to develop a credit scoring system in the most properway to minimize credit risk It also can help MSEs improve their performance andbusiness efficiency through enhancing key factors that have a great impact on theirbusiness

Determinants of loan repayment performance to help the bank separate good qualitydebtor and bad quality debtor is expected to come out as the result of the research

1.3 Objectives of the Study

1.3.1 General Objectives

The general objective of this study is to analyze a credit risk of determinants of loanrepayment performance of micro enterprises and household businesses in Vietnam

1.3.2 Specific Objectives

Specific Objectives of the Research is:

1. To identify determinants affect loan repayment performance of micro businesses in Vietnam Micro business includes micro enterprises and household business.

2. To identify association between categories of covariates and loan repayment performance.

3. To identify the key factors for credit risk of micro credit loan repayment that contribute on Creditors in Vietnam.

4. To investigate the most appropriate model for analyses of credit risk of micro business loan repayment in Vietnam Credit risk is the possibility of a loss resulting from a borrower's failure to repay a loan or meet contractual obligations (as by Investopedia)

1.4 Research Questions

From the above point of view and Objectives, research question problem as

indicated as follows

1. What are determinants on loan repayment performance of micro business

2. What are key factors affecting default rate of micro credit in Vietnam?

3. Which model is most appropriate for analyses of credit risk of microcredit loan repayment in Vietnam?

1.5 Scope of the Study

Research subjects are micro enterprises and trading household borrowers (MSEs) inVietnam MSEs is defined in term of capital amount, annual turnover, and number

of employees as per Vietnam Enterprise Law

These MSEs are located in Hanoi and neighboring area Hanoi is the capital ofVietnam with the highest density of enterprises and household businesses fromvarious fields and sectors in the economy With data and sample taken within 2continuous years of 2018 and 2019, from Hanoi and rural area – typical area, theresearch outcome is expected to indicate the overview of loan repaymentperformance of micro business in Vietnam

1.6 Structure of the research

With the aim at providing a full picture of micro business in Vietnam anddeterminants of credit risk of their loan repayment performance in the most detailand transparent way, this research is presented into 5 chapters as follows:

Chapter 1: Introduction – This chapter reveals basic ideas of the research in

terms of background, objectives, subject, and scope of the research, andresearch questions in order to provide readers with general view andmotivation of the Researcher

Chapter 2: Literature review – The 2nd chapter introduce and build

comprehensive understanding over both basic and complicated concepts inbanking and finance, explanation of credit risk, as well as a consolidatedreview of determinants of credit risk to the loan repayment performance ofmicro business, research gap as well as popular research models used inanalyzing these determinants in Vietnam and in the world

Chapter 3: Research methodology – After Literature review, the 3rd chapter

focus on a detail review of relevant models and the research model chosen toapply in this research Chapter 3 also describes how the research is designedand carried out, how data is collected, and cleaned before processing andanalyzing in Chapter 4

Chapter 4: Data analysis and findings focus on presenting and analyzing

data and results collected This chapter also aims at discussion of researchresults and comes up with relevant findings

Chapter 5: Conclusion – the last chapter summarizes the contributions,

recommendations as well as limitations, and intentions for future research

CHAPTER TWO: LITERATURE REVIEW

2.1 Concept and Definition

2.1.1 Overview of Micro Credit

Microcredit is “the extension of very small loans (microloans) to impoverishedborrowers who typically lack collateral, steady employment, or verifiable credithistory It is designed to support entrepreneurship and alleviate poverty”

Microfinance—also called microcredit—is a way to provide small business ownersand entrepreneurs access to capital Often these small and individual businessesdon’t have access to traditional financial resources from major institutions (Thebalance.com)

In almost all previous Research and journals, microcredit is understood as a verysmall loan amount given to the poor or business unit, or women who want to starttheir own microenterprises

The average loan amount differs from countries to countries By [34], in USA,average microcredit is US$9,732 In Malaysia, that is US$427, that is in India, that

is US$206

2.1.2 Micro and small enterprises (MSEs) and Household business (HHs)

and Concept of MSEs credit

The SME law in Vietnam has already been passed by the National Assembly andtook effect from 1st January 2018 As per the law, “SMEs are defined as micro,small, and medium-sized enterprises having no more than 200 employees registeredwith the state social insurance scheme in a year and meeting either of the followingtwo criteria:

 total capital shall not exceed VND100 billion (around USD4.4 million);

 total revenue of the preceding year shall not exceed VND 300 billion (aroundUSD13.2 million)”

Among that, Micro and small enterprises (MSEs) is the concept to refer to verysmall and micro enterprises, which total revenue of not over USD 01 million, which

is similar to VND22 billion, with an average number of employees of 50

By Decree No 56/2009/ND-CP, MSEs and HHs account for nearly 98% businessunits of the economy in Vietnam Their total capital normally not exceed VND10billion Their total working capital demand is around 1 billion per year and theirnormal tenor is from 3 months to 9 months, which equals to an average loan amount

of from VND 300 million to 900 million

These characteristics, compared with the concept of microcredit is different.Because microcredit is referred to a very small loan amount, including the poor.Hence, in this study, I would like to use the concept of MSEs credit, which isexplained in detail as below, to fit with research subjects in Vietnam in this research

MSEs credit

According to the internationally-used concept of Microcredit and regard to theobjectives of the research, a new concept will be created within the scope of thispaper MSEs credit is understood and defined as “credit provided to micro and smallenterprises and household business for their business to a certain amount andwithout collateral” MSEs credit includes loans provided MSEs, not included otherfinancial services

From now in this research, MSEs is understood both as very small and microenterprises and household businesses This concept will be used throughout theResearch and the main study objectives of this Research

MSEs in Vietnam and as defined in this Research, are classified into 3 main groups,

as per the practice of banks and experience of experts, as below:

 Startup enterprises with years in the business of fewer than 2 years They are

in their early stage and has earned the very first income Usually, they don’thave enough or can’t prove financial ability and don’t have collateral to get aloan from banks Desired loan ticket size is around 300 to 500 million VND

 Micro and small enterprises whose precursors are household for severalyears and seek for more capital to scale up their business, to move to highertarget in terms of revenue, benefit, as well as open more branches, etc…Desired loan ticket size is from 500 million VND to less than 1 billion VND

 Household or family business with a number of employees less than 10 people and average loan ticket size of 100 million VND

2.1.3 Characteristics of MSEs credit

Number of MSEs is very high in Vietnam, and most of them have demand to get aloan from banks and financial companies But just a small percentage can get access

to the capital they need, because of some characteristics of MSEs, such as: theydon’t have collateral, their cash flow is frequent by month, even by day Theirdemand is repeatable which means after paying the first loan, they will ask for anew loan They also can ask for a higher loan amount if they prove their goodpayment history after the first loan They don’t have a strong ability of managementand documentation which means the procedure for a loan should be as simple aspossible

Accordingly, the main characteristics of MSEs-credit includes, but not limited to:

 First, small loan amount and short-terms loan, usually up to VND500 millionand the term up to one year

 Secondly, payment schedules is equally monthly installment (EMI), which means MSEs pay an equal amount every month until the loan contract ends

 Thirdly, the loan procedure for MSEs credit is expected to be as simple aspossible, with least required documents as possible because MSEs inVietnam haven’t equipped themselves with an adequate management system

 Fourth, if the clients pay well and duly, they are eligible for a top-up loan with higher amount.

 Fifth, the interest rate is fixed during the loan term, because the term is up to

1 year (12 months)

 Lastly, MSEs loan are unsecured, which means no collateral is required Theloan is given basing on estimate of expected turnover and income of theclient, as well as other criteria related to the owner of the enterprise

2.1.4 Overview of Micro Creditors in Vietnam

In Vietnam, many years ago, banks and creditors mainly focus on big cooperate andmedium-sized enterprises Normally, these are secured loans with collateral as ofreal estate, factory, stock or payables Micro creditors are almost state-owned creditfunds, and the average ticket size is just around 10 million VND, which is verysmall and not enough for a MSEs’ working capital demand Normally, as experience

by experts in banking sector and standard by many Vietnamese joint-stock banks, aclient of MEs will need an average of VND500 million per year to finance theirworking capital demand

Several years ago, some foreign financial companies first appeared and made bigchanges to the banking and financial market in Vietnam, especially segments ofretail banking and micro-enterprises However, due to the sophistication in appraisal

of MSEs customers, as well as small NIM (net interest margin), just a few banksand financial companies offer loans to MSEs Normally, they just focus on retailbanking and ignore or pay little attention to MSEs In 2019, there are only 2 banksand 2 financial companies, and 2 P2P lending companies offer loan to MEs with themaximum amount of VND500 million

Lendbiz JSC is the only P2P lending company now in Vietnam can connectinvestors and MSEs for loans up to VND 1 billion In this study, secondary data wastaken from Lendbiz’s clients and credit officers

2.1.5 Credit Needs by MSEs in Vietnam.

As mentioned in the introduction part, the market of MSMEs in Vietnam is very big,includes both small, micro-enterprises, and household businesses This is estimated

to be around 6 million of borrowers including both MEs and HHs

The need for credit from MSEs is very high and increasing recently when theeconomy of Vietnam is in its fastest period and businesses are thirsty for capital.Main borrowers include, but are not limited to:

First, startup enterprises which need money to finance their business when they start

a business and are expected to have potentially good cash flow in the near future.Normally, these are enterprises that have operated for around 1 year

Secondly, enterprises that have been in business for a certain period of time but still

in their early stage and need money to extend their activity Normally, these areenterprises that have operated for 2-3 years

Thirdly, the biggest clients in the MEs group They already got loans from banksand all are secured loans They have certain amount of assets but that is not enough

to secure for loans at bank Hence, they seek for short term unsecured loan for theirworking capital demand Though the unsecured loan accounts for small percentageamong their total liability, but it still plays a very important part in their businessbecause it can help to maintain liquidity of the enterprise

2.1.6 Credit Risk Defined

According to Investopedia, “Credit risk is the possibility of a loss resulting from aborrower's failure to repay a loan or meet contractual obligations Traditionally, itrefers to the risk that a lender may not receive the owed principal and interest,which results in an interruption of cash flows and increased costs for collection [18]has similar findings Excess cash flows may be written to provide additional coverfor credit risk Although it's impossible to know exactly who will

default on obligations, properly assessing and managing credit risk can lessen theseverity of a loss Interest payments from the borrower or issuer of a debt obligationare a lender's or investor's reward for assuming credit risk.

 Credit risk is the possibility of losing a lender takes on due to the possibility

of a borrower not paying back a loan

 Consumer credit risk can be measured by the five Cs: credit history, capacity

to repay, capital, the loan's conditions, and associated collateral

 Consumers posing higher credit risks usually end up paying higher interest rates on loans”

There are several traditional ways to mitigate risk For instance, the lender/creditorwill take as many as possible information from the borrower, verify the informationand analyze clients using some tools depending on each type of client To secure forthe loan in case being turned into bad debt, most of the banks in Vietnam willrequire the borrower to take out their collateral, such as house, factory, car, etc orguarantees from third parties Also, by the rule of risk and return, banks also charge

a higher interest rate for loan with higher risk, as a way of compensation for thepotential loss they may incur

2.1.7 Relationship between Loan Default Problem and Determinants of

Loan Repayment Performance

Loan default problem is unwanted things and strategy to any creditors, from banks

to financial companies When Loan default occurs, it can knock down any creditors,destroy the lending capacity of even very wealthy creditors It erodes benefit ofbanks and also erode belief though banking and financial system Banks andcreditors will tighten credit policy and deny new applicants to get access to credit.Moreover, the loan default problem means the loss of future credit, loss of potentialdeposit from society

By [28], A loan default problem occurs when a borrower fails to make a payment ontime (defaults on a payment of interest or principle) or they do not comply with

any other agreement This is because the debtor is either unable or unwilling torepay In this research, Loan default problem means the failure of borrower in

making payments Non-defaulters are those who repaid the loan in due date and thedefaulters are those who didn’t repay the loan within the due date

2.1.8 Types of Loan

According to the Credit Institution Law of Vietnam No 47/2010/QH12 and Circular

No 39/2016- NHNN, Loan is the lending whereby a credit institution assigns orcommits to deliver a sum of money to customers for use for a specific purposewithin a certain period of time according to an agreement on the principle ofreimbursement Loan can be categorzied basing on loan tenor, loan purpose,etc Related to business loan, there are 2 main types of loans as below:

a) Short term loan for working capital demand: A typical business cycle of eachenterprise is normally less than 12 months, depending on business sector and scale

of each enterprise For instance, for an enterprise that provides service, it takes onlyaround 1 month For an enterprise that focuses on trading, it takes from 2-3 months.For a manufacturer, it may take longer from 3 to 5 months from the time when theybuy material, to the time they sell and collect money from the buyer The businesscycle is even longer with construction enterprises due to the specific characteristic

of this business sector It takes them 6-12 months, or even years to finish onebusiness cycle Accordingly, the demand for working capital is always short termand less than or equal 12 months

c) Medium and long term loan for investing of fixed assets: Besides demand forworking capital, enterprise also need medium and long term to buy or invest infactory building or fixed asset which is essential to medium and big enterprises.Depreciation time will be several years, and the payback time should be relevant.Therefore, medium and long term loans usually from over 12 months to 72 months(equal 6 years)

With the scale and characteristics of MSEs which most focus on service, trading orsmall manufacturing small manufacturing sector, the demand of big amount forfixed asset is not too much or they can well arrange with a short term loan In thisresearch, research subjective are short term loan under 12 months.

2.2 Related Literature Reviews to the Variables Used in the Study

In this section, related literature reviews of scholars on MEs - borrower’scharacteristics, Determinants of loan repayment performance, and Creditors’method in the appraisal of clients are addressed The literature review is made withthe objectives of identifying research/literature on micro-credit loan repayment ingeneral and determinant a credit risk of MEs credit loan repayment on microcreditand savings institution, in particular, understanding the problem and identifying anyresearch gaps in the area

Normally, key variables used as determinants of credit risks and loan repaymentperformance are divided into 3 main groups (Borrower characteristics, Businesscharacteristics, and Loan characteristics) which are: Gender, Age, Educational level,Residence, Marital status, Business Sector, Total sale, Total income, interest rate,loan amount, Number of dependents, extra income, an extra loan from other banksand creditors ,etc…

Key variables used as determinants of credit risks and loan repayment can becategorized as Qualitative and Quantitative data, regarding their own characteristics.Some popular and commonly used variables considered as key determinants of loanpayment performance in related Literature Review are reviewed as below:

2.2.1 Gender

Gender (Male/female) is a demographic factor that is proved to affect the behavior

of each person, in many areas, and loan repayment performance is not an exception

Female is considered more responsible and disciplined than male and is expected tomake loan repayment duly as the agreed indenture.

By [1] the researcher found that Gender factor had a significant effect with theregression coefficient value of 0.095 and a p-value of 0.0001 with a hazard ratio of1.10 The finding result shows that male clients had a default risk of 1.10 timesfaster than female clients That is, the risk of default was higher in men than inwomen It was because naturally female customers have a better willingness torepay the MSEs loan compared to male customers These results are in line with theresearch conducted by Rahman [11] and Syed Masud [12] which indicated thatfemale customers are easier to be affected by the pressure within a credit group andmore sensitive to the intervention or collection activities done by the creditor

In Vietnam, there haven’t been similar research on the effect of gender factor to loanrepayment performance but as many experts and experienced loan appraisers,female usually makes a good impression and feeling to creditors and is expected tomake the repayment duly because their characteristics are more gentle and exposed

to pressure –if any, which make them react more quickly with requirements ofpayment than male Moreover, Vietnamese women are believed to have the ability

to make savings and spend less money on entertainment activities than males, aswell as have better skills in cash flow management – qualitative criteria to manybanks and financial companies in Vietnam when they make a loan decision

2.2.2 Age

According to [1] Age was found to be a significant effect with the regressioncoefficient value of -0.024 and p-value of 0.0001 with a hazard ratio of 0.98

This finding indicates that:

Age of borrowers is expected to have the same direction, or positive impact on loanrepayment, which means, the older the borrower is, the less probability of defaulthis/her loan is It also means that younger clients would have 0.98 times faster of

loan default than the older ones This can be explained by a truth that older ageperson normally is wiser and more experienced in making a decision, opportunitycost of all options, so that the default risk is lower thanks to both positivewillingness to pay and ability to pay.

Also, [2] found that Age of the borrower’s was found to be as expected, as positiveimpact on loan repayment, which means as age increased, the probability of beingdefaulter is decreased

In Vietnam, there hasn’t been yet research on effect of age factor to loan repaymentperformance but as many experts and experienced loan appraisers, older age aremore likely to access credit from formal financial institutions than those withyounger age Because they are wiser and better in making decision, management oftheir life, their financial status which are essential in creating willingness to pay of adebtor Also, the older the debtor are, the possibility that they are married or havebeen responsible for someone else is higher, which means

For sure, we are mentioning borrowers who meet the requirement of banks in term

of age Internationally, people are within labor age are more eligible for requesting aloan, in term of both health condition as well as ability to make and maintainincome In Vietnam, only borrowers who are between 20 and 55 are accepted as apotential customer of banks and financial companies People over that age are notqualified for business loans, but retirement loan only

2.2.3 Marital Status

Marital status, includes 4 groups: single, married, divorced and widow However, interm of statistics and analytics for credit risk, it’s simplified into 2 main groups:married and not married

By [2], 78.94% of the borrowers who have defaulted problems are married, and24.49 % are not yet married From this, we may conclude that married borrowersare easier to have bad debts than unmarried ones According to [3], it also seems

true because married clients are likely to have larger families than the ones that arenot married Single clients are therefore more likely to pay off their debts than themarried ones.

However, this is not entirely true because in general, the owner of an enterprise,though small, with an average age of 30, mainly gets married Sample of MSEsborrowers is different from a sample of consumption borrowers By [3] it’s foundthat marital status had the regression coefficient value of -0.257 and p-value 0.0001with the hazard ratio of 0.77 The negative coefficient value means that thecorrelation between marital status and loan default problem was inverse direction.From a study conducted at a Tunisian MFI, [13] it’s also concluded that beingfemale and married is likely to increase the probability of loan repayment ofborrowers

Also, from some journals on the internet, it’s stated that married debtors wouldrepay their loan faster than the unmarried ones It is probably because the marriedones will be more responsible and mature as debtors

In Vietnam, marital status is used in credit scoring of some banks and financialcompanies, but just as reference information Just a few creditors such as FE Credit(a famous financial joint-stock company in Vietnam), HDBank (a famous joint-stock bank of Vietnam), etc use demographic variables in their credit scoringsystem, especially the marital status of MSEs owners, hence, no data or literaturereview of marital status in Vietnamese research or journals But, as experts in creditapproval, given the other conditions are the same, the married business owners have

a clearer commitment to making loan repayment

2.2.4 Educational level

Educational level refers to the highest level of schooling that a person has reached.This variable is very meaningful determinants in assessing credit risk, especially ofMSEs, because, to be the owner of an enterprise, a certain level of education and

experience is required Professional background, experience, and ethics of thebusiness owner will directly affect organizational culture as well as business results,usually in the same direction.

As usual, the Education level will be divided into 4 main groups: Primary,Secondary, High, University, and above However, depending on the research target,education level can also be categorized as illiterate, primary, etc…Within theresearch of credit risk of MSEs borrowers, high school, diploma, and undergraduateare used because there are almost no illiterate borrowers among MSEs’ owners

By [9], Using the education level of elementary school as the control variable, thefactor of education level had a significant effect on the default level on MSMEcredit at Bank XYZ By[2], it is shown by all dummy variables of education have ap-value of 0.0001 (less than the significance level of 0.05) The negative coefficientvalues on education level variable of junior high school, senior high school,diploma, and undergraduate show that clients with lower last education level had afaster default risk than the clients with higher

The educational level coefficient had a positive sign indicating that education has adirect relationship to the repayment rate and this showed that as level of educationincreases, borrowers enhance their ability to access, evaluate, and understand newproduction techniques The result of this study showed that the higher the literacylevel of the clients, the higher will likely be non-default

They found that education is an important determinant of loan repayment Aneducated client can use modern technologies, perform farming activities based oncropping calendar, and manage resources properly

Education level is expected to be one of the factors influencing credit repaymentbecause it affects one’s individual character The higher the education level, thewider the knowledge and insight owned

2.2.5 Residence

By [13], The scholar analysis of credit risk on MEs credit loans in Kenya usingsurvival analysis techniques, the variable of Residence is used because the area ofresearch is quite large Based on the results calculated, the region with the highestpercentage of default existed in Medan region with a percentage of 7% It becamethe basis why Medan was chosen as the reference region variable in the modeling

By [10], The results of Cox regression show that all dummy variables of the regionhad an effect on the level of default on MSME credit at Bank XYZ It can be seenfrom the p-value of 0.0001 (less than the significance level of 0.05) However, thismay not entirely true because big cities have more MSEs than rural areas and smallcities The competition level among MSEs and the demand for business loans in bigcities is also higher than that in small cities Accordingly, if we just focus on theratio of MSEs which got defaulted, it may not reasonable

In Vietnam, Hanoi and Ho Chi Minh city are the 2 biggest cities of Vietnam that arehome to many MSEs And loans for MSEs also focus on these 2 cities The data forthis research has been taken for the area of Hanoi and Ho chi minh city Besides, tomake it precise, many other data should be taken in parallel with the variable ofResidence Therefore, this variable will not be used in Research Instead, thevariable of Housing will be used Detail of this variable will be mentioned later

2.2.6 Distance of clients from institution

[1] found that if the nearer the borrower is to institution increases, the higher thepossibility of borrowers to repay their loan is However, [3] clearly showed that69.16% of the defaulter respondent’s residence and businesses were near HarariMEs credit institutions, whereas 30.84% were not near to Harari MFI As a findingresult, distance of borrowers from the offices doesn’t affect the loan repayment rate

of borrowers

These 2 conflicted findings make this variable not a good option as a keydeterminant of credit risk of loan payment performance

Moreover, nowadays, information technology has a great growth Just withinseveral months, or just several weeks, another technique may be born and replacedthe old one which was already very good For example, in Vietnam, number ofpeople who use internet for their banking transaction is quickly increased Thisfocus most on transferring money, saving money, etc Similarly, borrower do notneed to go to bank transaction office to pay their loan but can make it throughelectric wallet in mobile or their laptop, in a very easy and convenient way

Hence, distance of clients from institution is not really a good variable asdeterminants of credit risk of loan repayment performance It is just helpful whenthe loan already get overdue and come to a bad debt and require field collection Forprediction of the default risk, it’s not helpful and not used in this Research

2.2.7 Business sector

By [1], a survivor function curve for construction sector experienced the possibility

of faster default with the percentage of 23%, followed by transportation, storage,and communication sector with 19.4% and small manufacturing industry sector with19.2% [26] carried out a study in order to determinants of access to formal financialsources of micro and small enterprises by using a two binary logistic regressionsmodel According to his finding, “MSEs engaged in small manufacturing andconstruction sector easily access credit from banks and MFIs as compared to thoseengaged in trading and service sector”

The reason for this might be enterprises in the asmall manufacturing andconstruction sector tend to have more fixed asset which can be used as collateral forloan with big amount However, MSEs, especially in Vietnam, almost are businessunits in trade, service and light industry Their fixed asset is not at very big value.Even the owner havn’t had house of their own Therefore, this may not be true inVietnam

As experts in banks, and data from Agribank of Vietnam, small manufacturingsector is not a good sector to commercial banks because of poor loan repaymentperformance By unofficial data of this bank, the level of default for each sector aresmall manufacturing sector, small manufacturing , service industry, from high tolow, respectively The trade sector enjoys the lowest ratio of default This may comefrom a truth that the benefit ratio and cash flow of MSEs in trade sector aregenerally better than MSEs in heavy industry or MSEs in small manufacturingsector where business cycle is longer than that of trade sector Due to characteristics

of Vietnam, MSEs almost do not include small manufacturing sector because thatmainly taken care by big state owned enterprises

This variable is absolutely a key factor affecting default risk of borrowers and isused in this Research, and divided into 4 main groups: Manufacturing, Trade,Service and Construction

2.2.7 Collateral

Collateral is the asset that borrower use to secure for their loan In case they cannotpay the loan, the bank or the creditor will take that asset, sell it and compensate forthe loan and other cost that they have to incur from the defaulted loan

According to [2] it is revealed that, the causes of default that is identified were inadequateloans not requiring collateral Also it was revealed that there is a default problem as 80% ofloan officers agreed to that and there are delays on payment of loans

In Vietnam, the similar phenomena happen, loan repayment rate is significantlyrelated with type of collateral offered because collateral usually is house or car and

if the borrowers don’t want to lose their house, they will try their best to pay backthe loan However, as mentioned, as Vietnamese expert, collateral does affect thewillingness to pay of borrowers and probability that banks can get money backthanks to the seize of borrowers’ asset MSEs usually don’t have big value asset oftheir own Even he MSEs owners don’t have valuable assets to secure for the loan

Hence, in this research, loans are considered unsecured which there is no collateralrequire by creditors Accordingly, this variable is not used in the research.

2.2.8 Amount of loans

According to [2], [10] it is suggested that credit/loan size was found to negativelyinfluence the borrower’s loan repayment performance and loan size was negativeand significant determinant of the probability of loan repayment

This can be true or false in two ways: if the loan amount is suitable with capitaldemand and cash flow of the borrower, then it is the best option despite of big orsmall amount If the loan is not suitable, larger or smaller than borrower’s demandand cash flow, this will be very dangerous because borrower may use the capital inthe wrong way and, or cannot make enough money to pay for the creditors

After investigating, it’s found that loan amount is calculated according to demandand financial statement of the borrowers Many loans of small amount still getdefaulted and many loans of big amount do not get defaulted, so this variable is notchosen to investigate as one of factor influencing loan default

2.2.9 Repayment period

Repayment period equals loan term which means the total months from the time ofdisbursement to the time the borrower has to pay all the loan amount Loan period isthe due time for the debtors to repay the primary installment along with the interest.Loan period will affect the amount of installment and interest that will be paidevery

According to [3] The results show that when the average loan period increased, theclients who repay the installment in a long period would have a 1 time faster defaultrisk than the ones who repay in a short period

This is true in terms of risk management, because the longer the period is, the riskerthe investment (loan is understood simply an investment where interest is the extra

value earned for the investor) is This can be easily understood because when timechanges, there are many events, force majeure incidents that may happen and affectbusiness activities For i.e, Covid has made many MSEs go bankrupt, manyemployees lose their job and can’t pay their debt, etc.

However, MSEs credit in Vietnam almost is short term loans which finance workingcapital demand of MSEs Short term loan is loan with repayment period under orequal 12 months Therefore, Repayment period is not a necessary variable and used

in this research

2.2.10 Interest rate

[11] found that, the causes of default that are identified were inadequate highinterest rate This means that if a borrower has to pay too much for a loan that hetakes, the probability of repayment is low It can be simply explained like this: anMSEs borrow money to put into his business, to earn benefit at a certain rate, let’ssay 20% If it has to pay an interest rate of 30%, which is higher than the benefitratio that the money it borrowed can make, the cash flow will be inadequate andthere will be no money to pay for the creditor

[2] also use interest rate is one of variables to analyze the probability of default risk.However, in this Research, and in Vietnam, a short term loan is given to borrowerswith fixed interest rate, which means all MSEs with given qualifications will enjoysimilarly In a few Vietnamese banks and financial companies such as FE credit,MSB, interest rate isn’t used as criteria in their scoring system Hence, this variable

is not used in this research

2.3 Cox Regression Model

To analyze credit risk in general and default ratio, researchers can use manydifferent Models, from non-parametric models such as Life tables, Kaplan Meier, tosemi-parametric models such as Cox Regression Model, or other more complicatedparametric models Among these models, Cox regression model is known as a

useful tool for survival analysis which is often used in the medical area to define thetime that patients may get covered from the disease, or estimate the time durationuntil an event Recently, the application of Cox Regression Model is more popular

in many other fields, including risk management and credit risk modeling

By [7], Cox Regression Model is a methodology created by Doctor Cox, whichhelps us to predict the probability/outcome of some event which can happen at acertain point of time, instead of just providing outcome (yes/no, 1/0, etc…) of someevent as normal Linear Regression

By [15], when comparing between Logistic Regression Model and Cox HazardModel, it can be seen that the two models have a very similar performance in theGini coefficient on the development (training) and comparison (validation) sample,but the Cox model shows a little more stable Moreover, Cox model also helps topredict time to default, which is not available in other models

Table 2.1 Comparison of models on the random sample

Summary

Logistic regression Gini

Cox model Gini

Logistic regression lift 10%

Cox model lift 10%

Also, by [17], Cox’s proportional hazards modeling is applied in the generation ofthis model since it’s the most suitable for survival data when proportional hazardshave been proved in various groups Therefore, to predict the probability of LoanDefault – a situation that a debtor can’t make the repayment duly at a given timeduring the loan period and the time to default as well as the effect of covariates, CoxRegression Model is the most suitable tool and is used in this Research By [19], it’sagreed that Time to default is key factors in credit scoring

Also, by [21], different from Kaplan Meier model which only allows 1 variable,Cox’s regression allows several variables to be taken into account and tests theindependent effects of these variables on the hazard of the event

2.4 Research Gap

2.4.1 Gap in Research Model

Up to now, in Vietnam, there are several research on credit risk and credit scoring,but no research on the similar topic using Cox Regression Model Binary regressionmodel is used in some research, but this model provides only Yes/No result on theprobability of default

The Cox Regression procedure is useful for modeling the time to a specified event,based upon the values of given covariates One or more covariates are used topredict an event, such as failure to pay, time to default, etc… The central statisticaloutput is the hazard ratio Data contain censored and uncensored cases Similar tologistic regression, but Cox regression assesses the relationship between survivaltime and covariates

To deeply dig into this topic and come out with the most proper determinants to loanrepayment performance or default ratio, this research will use the Cox RegressionModel which can predict both hazard ratio of default and time to default

2.4.2 Gap in Variables

Variables in Section 2.2 are the most commonly used variables by internationalresearch Basing on characteristics of Vietnam economy and practice, as well asobservation by experts in banking and finance sector, 6 following variables will beadded to the list to investigates determinants which affect loan repaymentperformance, including Years in business, Percentage of share owned by owners, thestability of turnover in 6 latest months, Availability of Digital Sale channel, Housingand Number of clients Details of new added variables are as below:

2.4.2.1 Years in business:

Number of years that MSEs and its owners have worked in the field, to the timeMSEs – same business sector, apply for a loan This variable is used in the Researchbeing observed from the fact that the more experience the owner has in the sectorthat his enterprise has business in, the higher probability the borrower can repay theloan It is explained that if the owner has more experience, he or she will have alarge network with people in the field, mindset, and experience to build a suitable,smart, and realizable strategy and execution plan for his/her enterprise He or shealso understands his/her position and competitors in the segment that they focus on.Actually, in Vietnam, this variable is used by quite many banks and financialcompanies when they make credit assessments to offer a loan or not to borrowers,not only individual but also enterprise customers However, these creditors haven’tapplied this variable into their scoring system or model, at least as searchabledocuments and unofficial information by experts in the field Similar research andjournals on the effect of this variable to loan repayment performance are notavailable currently

2.4.2.2 Percentage of ownership

Percentage of share that the owner has in the company This is a variable to measurethe influence of the main owner to the enterprise – or MSEs borrowers In small andmicro enterprises, and household business, personality, and background of the owner strongly influence to strategy, efficiency and culture of the MSEs he or she own

There are two types of private enterprises in Vietnam: limited liability company(Ltd) and joint-stock company (JSC) and these 2 types of companies allow manyshareholders to put money in the company The higher the number of shareholders

is, the less percentage of share each shareholder owns, in general However, theshareholder who takes the highest percentage of share, usually also the founder, will

try to own more than 50%, to ensure that he or she will seize all activities of thecompany and influence all important decisions to his or her enterprise.

In Vietnam, VPbank and Lendbiz have regulated that only MSEs in which the mainowner owns more than 50% of share is eligible to apply for an MSEs loan Thiscomes from their experience during years giving loans to this type of borrowers, aswell as experience in collection Up to now, there is no research announced byVbank and Lendbiz on the outcome of the mentioned policy In this research, authorinvestigate the influence of this variable, with two relevant values: under and equal50%; over 50%

2.4.2.3 The stability of turnover in 6 latest months: Measurement of the stability

of MSEs’ turnover within 6 latest months

For young enterprises such as MSEs, it’s not very easy to maintain a stable turnoverflow and cash flow accordingly They have to try their best to adapt with thecustomer and market, learn to survive, and grow in a competitive market If theirturnover is not stable, it may lead to the loss of liquidity and hence, MSEs will fail

in making loan repayment Experts in Vietnamese banks and author’s ownexperience, this is an important factor Though several banks in Vietnam also usethis as a requirement to filter qualified clients, they haven’t made a scoring systemwith this factor and make the measurement of how it affect default risk ratio

In this Research, there are 2 value for this variable: difference > 30% compare tomean value, difference ><30% compare to mean value

2.4.2.4 Availability of Digital sale channel: measurement if MEs use digital sale

channel besides traditional sale channel (Direct sale)

This is expected to heavily affect their business in the era of technology In this era

of 4.0 technology Revolution has changed our life and the way we live and work.People don’t need to go to see each other frequently They can also stay at home andwork, make transaction, do shopping, etc…This is even more clearly when Coviddisaster came and people have to activate the mode of work from home

Housewife do shopping through e commerce and make payment by e-wallet Then

if MSEs just have a traditional sales channel, they can’t catch up with this changeand quickly lose their market share to competitors who are pioneer and already set

up digital sales channel This can be easily seen through the success story ofAmazon Retailer or Aeon Mall, who have survived and get amazing growth rateover other traditional retailers

In the research, there are 2 value for this variable: Yes/No

2.4.2.5 Number of clients: Number of clients that MSEs has at the time it applies

for a loan

The more clients it has, the less dependent on a few big clients, which means therisk is diversified from late payables or decrease in turnover due to the withdrawal

of a certain number of customers

This is strongly agreed by appraiser and approver from banks, as well as debtcollection experts

2.4.2.6 Housing

Owners of MSEs almost does not have a house or valuable house of their ownbecause of young age Also, many of them put all money they have into the businessand live in a rent house This is very normal in Vietnam

However, if they already have a house or in the process of buying a house for theirown, things are different This means they already have certain success and can earnmoney from the business

This variable has two value: Owned house and Rent house which can influence andindicate the ability to make repayment of borrowers

CHAPTER THREE: DATA AND METHODOLOGY3.1 Description of the Study Area

The study was conducted in the capital of Vietnam - Hanoi Hanoi is located in theNorth of Vietnam and has been chosen because it is a typical city with many types

of enterprises and household businesses with an estimated population of 8.1 million

as of 2019 with 12 districts, 17 rural districts and 1 town

According to a ranking recently by PricewaterhouseCoopers, Hanoi has the highesthuman development Index among the cities in Vietnam and will be the fastest-growing city in the world in terms of GDP growth from 2008 to 2025 The relevantdata is collected through Lendbiz– a P2P lending company mainly and VPBank -partially with borrowers doing business in Hanoi and neighboring areas

3.2 Data Description

The data used in this study was both quantitative and qualitative data types Thedataset was collected from secondary data sources of Lendbiz and VPBank.Secondary data was collected from reports, unpublished documents, and datasources of 200 borrowers who took loans beginning from January 2018 and endingDecember 2019 with loan term of 10 months to 12 months

The data consisted of uncensored data (customers who had a status of repaymentfailure until the end of this study period) and censored data (customers whosecredits had been paid off and customers whose credit had not been completed butstill had good repayment status until the end of the study period

The data set was contained the client description variables, personal informationabout the MSEs’ main owner (age, gender, educational level, marital status,housing, years in business), the MSEs itself (business sector, availability of digitalsale channel, number of clients, the stability of turnover within 6 latest months) Thedata was collected by appraisal credit officer and coded, cleaned It was