Households access to informal rural credit an analysis from VHLSS 2008

This paper explores determinants that influence loan amounts borrowed by rural households from the informal sources using the Heckman two-step model and the data in the Vietnam Household

VIETNAM - NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

HOUSEHOLD ACCESS TO INFORMAL

AN ANALYSIS FROM VHLSS 2008

By

LE ANH THU

MASTER OF ARTS IN DEVELOPMENT ECONOMICS

HO CHI MINH CITY, August 2012

UNIVERSITY OF ECONOMICS INSTITUTE OF SOCIAL STUDIES

VIETNAM - NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

HOUSEHOLD ACCESS TO INFORMAL

RURAL CREDIT:

AN ANALYSIS FROM VHLSS 2008

A thesis submitted in partial fulfilment of the requirements for the degree of

MASTER OF ARTS IN DEVELOPMENT ECONOMICS

This thesis would not have been accomplished without the kind assistance and enthusiastic guidance of several individuals who have in one way or another contributed toward to the formation and fulfillment of this paper

First of all, I would like to express my utmost gratitude to Dr Nguyen Trong Hoai and Dr Pham Khanh Nam for sharing their practical profound insights and steadfast encouragement in term of theoretical literature and techniques that are helpful to this study

I would like to express my special thanks to Dr Tran Tien Khai and Dr Ha Thuc Vien who have inspired initiatives and passion of rural credit in me to carry out this interesting study

I would like to express my deep and sincere gratitude to Dr Le Van Chon for his scientific guidance, patient encouragement and invaluable advice, which he has provided throughout the time of preparation and accomplishment of this paper

I sincerely would like to thank all my loved classmates in class MDE17 and staff in the VNP office, who always give me their restless assistance when I was in trouble; especially Nguyen Thi Thuy Thanh, Dinh Thi Thu and Phan Thach Truc for creating motivation for me due to their brilliant talent

I would like to express my special massive thanks to my bosses MA Nguyen Duc The and Mrs Van Thi Le who have given me opportunities and spiritual assistance to accomplish this study

Last but not least, I must express my most gratitude to my parents and my wife, Bui Thi Thanh Thuy, for their enduring understanding and spiritual assistance which help me overcome the hard times

The purpose of this paper is to examine household access to informal financial markets in Vietnamese rural areas It applies the theoretical framework of asymmetric information to investigate the rural credit markets in Vietnam

This paper explores determinants that influence loan amounts borrowed by rural households from the informal sources using the Heckman two-step model and the data in the Vietnam Household Living Standard Survey which was carried out in the year 2008 It is found that household expenditure, household assets and number

of working earners strongly and positively impacts on probability of credit access; gender of household head highly negatively influence credit accessibility Notwithstanding, informal loan appears almost in household’s activities for both consumption and production purposes, they still mainly rely on relationship of friends and relatives rather than fully make use of informal credit institutions and other sources

Key words: Informal rural credit; cross section; Heckman model; VHLSS 2008;

Vietnam

ACKNOWLEDGEMENT i

ABSTRACT ii

CHAPTER 1: INTRODUCTION 1

1.1 Problem statement 1

1.2 Research objectives 3

1.3 Research questions 3

1.4 Justification of the thesis 4

1.5 Organization of the thesis 4

CHAPTER 2: LITERATURE REVIEW 5

2.1 History and origin of the rural credit 5

2.2 Theoretical review of rural credit markets 6

2.2.1 The theory of monopoly 7

2.2.2 Asymmetric Information 9

2.2.3 Indirect screening mechanism 9

2.2.4 Direct screening mechanism 10

2.3 Empirical studies 11

2.4 Conceptual framework 19

2.5 Chapter summary 22

CHAPTER 3: DATA AND RESEARCH METHODOLOGY 24

3.1 Background of rural credit markets in Vietnam 24

3.2 Data 28

3.3 Research methodology 29

3.3.1 Descriptive analysis 29

3.3.2 Econometric model 29

3.4 Chapter summary 32

CHAPTER 4: ANALYSIS RESULTS 34

4.1 Rural credit markets 34

4.2 Intention of loans 35

4.3 Different loan amounts analysis 36

4.4 Descriptive statistical analyses 38

TABLE OF CONTENTS

4.4.2 Bivariate analysis 39

4.4.2.1 The nexus of loan amount and age of household head 40

4.4.2.2 The nexus of loan amount and gender of household head 41

4.4.2.3 The nexus of loan amount and education of house head 41

4.4.2.4 The nexus of average loan amount and number of dependents 42

4.4.2.5 The nexus of average loan amount and number of adults 43

4.4.2.6 The nexus of average loan amount and expenditure 44

4.4.2.7 The nexus of loan amount and total value of asset 44

4.4.2.8 The nexus of average loan amount and expense on livestock 45

4.4.2.9 The nexus of average loan amount and cultivated areas 46

4.5 Empirical results 47

4.6 Chapter summary 51

CHAPTER 5: CONCLUSIONS & POLICY IMPLICATIONS 53

5.1 Conclusions 53

5.2 Policy implications 54

5.3 Limitations and directions for further studies 55

5.3.1 Limitations 55

5.3.2 Directions for further studies 55

REFERENCES 57

APPENDICES 60

Table 2.1: Early empirical studies on rural credit 15

Table 3.1: Sample size by region 22

Table 3.2: List of variables 32

Table 4.1: Status of household’s access to rural credit markets 34

Table 4.2: Loan distribution over financial sources 35

Table 4.3: Loan distribution over loan intention 36

Table 4.4: Different amount of loan on rural markets 37

Table 4.5: Different loan amount on the informal sector 38

Table 4.6: Univariate analysis 39

Table 4.7: The Heckman two-step model for informal loan amount 47

LIST OF TABLE

Figure 2.1: The vicious circle of poverty 5

Figure 2.2: Unorganized financial markets in LDCs 7

Figure 2.3: Monopoly & competitive markets 8

Figure 3.1: Rural loan in the formal sector 25

Figure 3.2: Rural loan in the semi-formal sector 26

Figure 3.3: Rural loan in the informal sector 27

Figure 3.4: Sample size 29

Figure 4.1: Gender of household heads 39

Figure 4.2: The nexus of average loan amount and age of household head 40

Figure 4.3: The nexus of average loan amount and gender of household head 41

Figure 4.4: The nexus of loan amount and education of household head 42

Figure 4.5: The nexus of average loan amount and number of dependents 43

Figure 4.6: The nexus of average loan amount and number of adults 43

Figure 4.7: The nexus of average loan amount and expenditure 44

Figure 4.8: The nexus of loan amount and total value of asset 45

Figure 4.9: The nexus of loan amount and expense on livestock 45

Figure 4.10: The nexus of loan amount and cultivated areas 46

LIST OF FIGURES

CCF: The Central Credit Fund

LDCs: Less developed countries

GSO: The General Statistics Office of Vietnam

HH: Household

ROSCAs: Rotating Savings and Credits Associations

OLS: Ordinary Least Squares

PCFs: The People’s Credit Funds

SBV: The State Bank of Vietnam

VBARD: The Vietnam Bank of Agriculture and Rural Development VBP: The Vietnam Bank for the Poor (formerly)

VBSP: Vietnam Bank for Social Policies

VFA: The Vietnamese Farmers’ Association

VHLSS: Vietnam Household Living Standard Survey

VWU: The Vietnamese Women’s Union

WB: The World Bank

ABBREVIATION

CHAPTER 1: INTRODUCTION1.1 Problem statement

The absence of formal credit sources for the rural household is a critical constraint to agricultural development in many developing countries It is due to information asymmetry in the relationship of lenders and borrowers leading to problems of adverse selection and moral hazard (Stiglitz & Weiss, 1981), (Jung, 2000); the consequence is often that the rural lending is perceptibly undetermined Moreover the imperfect information appears in many government policies supporting credit for rural households The formal credit does not satisfy demand for financing projects or expenditure (George & James, 2010) In addition, formal institutions respond late to small loan amount (Lipton, 1976) Therefore, borrowers seek credit in the informal sector (private moneylender, parents, relatives, friends, etc.) to satisfy their production and consumption (Pham & Izumida, 2002), Mpuga (2010)

The poor household in rural areas inevitably relies mainly on the expensive informal credit to finance their production, expenditure, tuition fee, wedding, funeral etc; for that reason informal credit plays a crucial role in rural development with evidences from (Adams & Graham, 1981), (George & James, 2010)

The operations of informal financial sectors are concerned with many diversified aspects including the supply of rural formal credit market is not able to satisfy the growing demand of rural households and many borrowers turn to informal credit markets for their production and consumption needs (Takashi, 2009); almost characteristics of the rural farm households are risky and uncertainty, usually dealing with small scale of debt, while the supply of the formal sector often concentrates on large scale projects (Barslund & Tarp, 2008), (Takashi, 2009); transaction cost of the formal credit is higher than the cost implied so informal lenders provide financial services with the competitive cost (Adams & Graham, 1981), (Guirkinger, 2008), (Mpuga, 2010)

Rural farm households, especially in LDCs, do not generate enough affordable finance to break the vicious circle of poverty One of the factors constituting key obstacles could be that agricultural product activities are highly dependent upon weather and climate conditions; and there is few credit institutions reaching the rural poor In Vietnam, there are about 70 percent of the population still lives in rural areas and depending on agriculture mainly for their livelihood (FAO & Shoji, 2011) A renovation for transforming from the agricultural sector to a more modern society remains a challenge for Vietnam To get promotion of agricultural production and transformation is to start from how to access credit for poor households who lacking of collateral, educated and information is a key ingredient;

it is also the first objective of any poverty oriented strategies for the rural development of the financial system (Barslund & Tarp, 2008)

In recent years, Vietnam has been launching many programs to help the poor increase income by providing credit services for them There are two prime credit sectors coexist in finance market (Barslund & Tarp, 2008) comprising the formal sector which usually is government-subsidized and charges relatively low rates of interest (Hoff & Stiglitz, 1990): The Vietnam bank of Agriculture and Rural Development (VBARD), The Vietnam Bank for the Poor (VBP), The Vietnam Bank for Social Policy (VBSP), The People’s Credit Funds (PCFs), The Central Credit Fund (CCF) and other private and commercial bank involved in rural lending (Takashi, 2009); and the informal sector: Group Lending, Moneylenders, Rotating Savings and Credits Associations (ROSCAs), relatives and friends (Pham & Izumida, 2002) which plays an important role for farmers and generally out of their own equity (Hoff & Stiglitz, 1990) The purpose of borrowing from informal sources is not only for production but also for expenditure and other unexpected events Beside those credit sectors, the poor farmers still access to credit markets hardly due to lacking of collateral and assets Semi-formal credit sector appeared then as a powerful instrument aiming to alleviate poverty by lending a group of borrowers with low interest rates The Vietnamese Women’s Union (VWU) and the

Vietnamese Farmers’ Association (VFA) are the two largest Vietnamese mass organizations in the semi-formal sector (APEC, 2011)

Although many financial sectors coexist in the Vietnamese rural credit market, households have a tendency to get loan from informal markets with the usury interest rate since almost each of household only has a small area for cultivation; furthermore, agriculture is usually affected by nature; due to high transaction costs and production risks, the banks and other credit organizations tend to avoid to give a loan to individual farmers (Takashi, 2009) For those problems, this paper attempts

to examine the determinants of household credit access to the informal sector by using the data of VHLSS 2008 and the Heckman two-stage model

1.2 Research objectives

The specified objectives of the study are to

(i) Identify the determinant factors that influence the credit access of households in the informal credit market in Vietnamese rural areas

(ii) Draw some policy implications in order to improve household’s access to the informal rural sources

1.3 Research questions

This study was intended to deal with the main research question is “What are the determinant factors that impact the amount of loan in the informal sources in Vietnamese rural areas?” In order to facilitate answering the main question, I divide

it into two sub-questions (i) Do productive assets of household head positively and significantly impact loan amount in the informal sector? and (ii) Does productive capital of household head negatively and significantly impact loan amount in the informal sector?

The scope of the study will be limited in using the data set of VHLSS 2008

to examine determinant factors of rural household access to credit markets in Vietnam

1.4 Justification of the thesis

This thesis attempts to fill the gap of lacking selection bias when using the VHLSS data set to estimate determinants of household credit access

This paper also tries to update new information and changes of influences on household access to credit in Vietnamese rural areas by using the data set of VHLSS 2008

The evidence from the paper is informative and useful for rural development authorities Therefore they are able to have a general vision on rural credit markets

in Vietnam and then propose some practical policy recommendations to develop Vietnamese rural credit markets, alleviate the rural poverty and also improve rural livings as well

1.5 Organization of the thesis

The rest of the paper is organized into four more chapters Chapter 2 presents history and origin of rural credit, theoretical review of rural financial market, and empirical studies which were carried out inside and outside of Vietnam Chapter 3 describes rural credit market in Vietnam, data, research methodology and analytical framework Chapter 4 analyses the empirical results, identifies determinants of household credit access and gives some quantitative analysis of those factors Chapter 5 concludes, suggests some practical policy implications; limitation and direction for further studies are also discussed in this chapter

CHAPTER 2: LITERATURE REVIEW

In this chapter, some theories and studies of informal rural credit access, particularly those with emphasis on small farmers, are reviewed The review covers some experiences of developing countries and relevance aspects to the Vietnam situation The chapter comprises a description of the history and origin of the rural credit system; types and sources of informal rural credit; imperfect credit market on credit access and indirect and direct screening Theoretical literature of monopoly and imperfect information paradigm are also reviewed for basically analyzing Empirical studies about determinant factors of household’s access to credit are also described

2.1 History and origin of the rural credit

According to Heidhues and Buchenrieder (1999), the concept of agricultural credit was known in the seventeenth century, when Chinese farmers use rural credit

in agricultural production to increase their cash income, and to improve their living standards Similarly, in Western countries, the Germany Landschaften was founded

by Frederick the Great in 1769 and its principles were used by the system for the Federal Farm loan of the United States The original credit concept stems from the need to break the vicious circle of poverty, as shown in the diagram below

Figure 2.1: The vicious circle of poverty (Heidhues & Buchenrieder, 1999)

Source: Heidhues & Buchenrieder (1999), Rural financial market development

Low Consumption Low Income

Low Investment

Low Savings Low Productivity

The figure shows that poverty is affected by income, savings, investment, and productivity A low level in any of these factors will affect significantly poverty rate It is argued that the role of the credit program is to break this cycle, leading to

an increase in per capita income and thus increase the savings rate, investment rate and productivity (Heidhues & Buchenrieder, 1999)

2.2 Theoretical review of rural credit markets

Generally characteristics of the financial system in most LDCs is a state of

"financial dualism" consisting of organized and unorganized, they co-exist and operate parallel to each other The organized sector includes the commercial banks, the cooperative banks, credit societies, the rural banks, and government and semi-government agencies The unorganized sector mainly includes money lenders, indigenous bankers, rural development banks, pawnbrokers, landlords, merchants and traders Although specialization and division of labor have developed in the organized sector which is amendable to financial control, the unorganized sector is not (Ghatak, 1995)

The unorganized money markets in LDCs are quite complex and heterogeneous Figure 2.2 shows a typical picture of the unorganized money market

is shown There is a large number of non-professional groups of people joining in the rural sector to provide credit Two major source of credit are moneylenders & indigenous bankers; there is a variety of other sources which can be quite important

in some LDCs

Figure 2.2: Unorganized financial markets in LDCs

Source: Ghatak (1995), Monetary economics in developing countries 2nd

In developing countries, it is suggested that there are four competing theories

of rural credit markets (Hoff & Stiglitz, 1990) including (i) monopoly power of village moneylenders, (ii) imperfect information, (iii) indirect screening mechanisms, and (iv) direct screening mechanisms

2.2.1 The theory of monopoly

In this view, the village moneylenders in the informal market is characterized

as a monopolist and usually charge with extortionate interest rate to maximize profits with assumption of no competition among moneylenders (Hoff & Stiglitz, 1990) When the lender receives information from various sources, there is asymmetric information between lenders It is expected that some lenders have the advantage of lending to certain people This gives strength to the loan market in a segment of the market, where he was better informed than any competing lender When a lender loans is the best information, or lenders to provide loans only in the region, the lender is also likely to have enforcement power Even in the absence of collateral and the threat of physical punishment, a lender can only make sure that

UNORGANISED SECTOR

Rural development banks

Regional rural banks Professional

Pawnbrokers Merchants Traders Others

relatives Landlords Merchants Agriculturists Others

any borrower default on a loan is excluded from future credit A lender can use the lending strategy of pure monopoly to people when there is no problem of information and enforcement

The existence of monopoly profit or usurious interest rate can be illustrated with the help of a simple diagram (see figure 2.3)

Figure 2.3: Monopoly & competitive markets

Source: Ghatak (1995), Monetary economics in developing countries 2nd

Private moneylenders would persist to be a valued credit source in remote areas for a longer time due to their easier informality, approachability and flexibility Village lenders do not get involve much in the form of administration and information cost They are taking risks from lending to people who are not creditworthy, maintaining idle cash balances to provide for immediate credit needs These reasons could explain a part of high interest rates in the informal sector (George & James, 2010) The monopoly view cannot account for simultaneous existence of formal and informal sectors, despite the fact that interest rates of formal

sources are essentially below those charged in the informal sources (Hoff & Stiglitz, 1990)

Adverse selection exists when borrowers differ from respect to the probability of repaying their loan In this case, the terms of a credit contract may affect the average quality of loan applicants For instance, low risk borrowers may drop out of the market if banks raise the rate of interest (Bester, 1987)

Next to adverse selection, moral hazard also has received a lot of debates on credit markets in many developing countries Moral hazard appears when an individual attempt taking actions to maximize her/his welfare by causing other's detriment in situations where problems of information asymmetry prevent assignments of personal damage caused by her/his action According to Bester (1987), moral hazard involves loan contracts which influence directly to the behavior of borrowers Therefore higher interest payments may induce the borrower

to seek investment projects with a higher profit also riskier

Informational asymmetry, high transaction costs and uncertainty are specific characteristics of informal credit markets These characteristics typically lead to problems of adverse selection and moral hazard In order to overcome those problems, indirect and direct screening mechanisms are applied in lending procedures to expose quality of borrowers

2.2.3 Indirect screening mechanism

According to Hoff & Stiglitz (1990), Interest rate is considerably as an indirect mechanism which replies on the design of contracts by lenders The lender obtains information about the riskiness of the borrower, and induces the borrower to

take actions to reduce the probability of default and to repay the loan when he responds to these contracts in his own best interests A lender charges an interest rate which could influence directly on the riskiness of the loaning pool by either affecting borrower's actions (the incentive effect) or ranking potential borrowers (the effect of adverse selection) An interest rate, on account of either of these effects, is set too high This action inevitably leads to increase the riskiness of the applicant Therefore, an optimal interest rate maximizing the returns of lenders will

be set, which is a great deal of returns competition compared to the market-clearing rate In addition, also due to the influence of the interest rate on the nature of transactions, it is considered as an indirect screening mechanism, categorizing the riskiness of borrowers Those who are considered as worse risks are the ones willing to pay higher interest rates to get loan since they are fully aware of their low repayment probability (Stiglitz & Weiss, 1981)

There are two functions of interest rate, which serves as the function of a price and an indirect screening or incentive mechanism Whether there is competition or monopoly in the market, these indirect mechanisms are equally applicable

There are two other indirect mechanisms are characterized as incentive effects of terminations and market interlinkages to reduce the riskiness that borrowers undertake the actions desired by lenders First, the lender may use threat

of cutting off credit to stimulate desired borrower behavior; and the other is that lenders who are landlords or merchants may use the contractual terms in these other exchanges to affect the likelihood of default

2.2.4 Direct screening mechanism

In order to minimize cost of imperfect information, lenders can exploit particular direct screening mechanisms to monitor borrower’s behavior and enforce loan repayment in the credit market, withdrawing credit if the terms of the loan appear to be violated In developing countries, costs of direct screening, monitoring and enforcing loan repayment vary among lenders, which may lead to credit

rationing in credit markets Village lenders could take advantage of being a product of living near the borrower or being part of the same kinship group or a party to some other transactions to reduce significantly the costs of screening, monitoring and enforcement (Hoff & Stiglitz, 1990)

by-Interlinked credit contracts are also another extensively used mechanism The borrower and lender jointly participate in a market which is not really a credit market This factor would reduce the cost of contract enforcement from the lender, and raise creditworthiness of borrowers to the lender Therefore, problems of enforcement, incentive and screening are alleviative since transactions of those markets can directly interlinked to loans In several markets, when an economic relationship requires transactions, a greater surplus can be gained from this scope Lenders/traders tend to utilize interlinked credit contracts by entailing their clients promise to sell all their farming goods from the harvest to them The form of interlinkage is widely used in Thailand, Pakistan, India and among other countries (Floro & Yotopoulos, 1991)

2.3 Empirical studies

+ International empirical studies

Okurut et al (2005) attempted to identify determinants which influence demand for informal credit by using the Heckman two-stage model Data were collected from 10692 households surveyed in the Uganda National Household Survey 1999/2000 He found that determinants which influence positively and significantly informal credit demand are age, dependency ratio, education level, household expenditure, gender and regional location On the other hand, factors consisting of age, asset values, regional location and gender are negatively and statistically impacted in term of the supply side

To get more understanding why farm households prefer getting a loan from informal markets even if the formal loan is more expensive, Guirkinger (2008) collected a panel data set and employed the logit model with fixed effects & random effects to examine why rural households tend to have more loans in informal

sources in Piura, Peru Data was surveyed in two periods in the rural coastal area of Piura, Peru’s north coast First, there were 547 households surveyed in 1997 Second, 499 households were re-interviewed in 2003 He found that transaction costs in the formal sector are higher than in the informal sector It could lead to the informal effective cost of a loan cheaper; hence, it could be affordable for households access to informal loan in term of effective cost In addition, informal lenders have a more efficient mechanism in direct screening, monitoring and loan recovery techniques compared to formal lenders They can give loans to borrowers with more flexible & convenient conditions Thus risk adverse households tend to borrow from informal sources

An interesting paper of Akudugu et al (2009), he identified the characteristics that influence access to credit by women borrowers in Ghana Data of the research were obtained from the survey of 200 women farmers in the Upper East region of Ghana He found that rural formal credit has been in favor of women with the proportion of 44 percent compared to men is 56 percent Membership to economic associations, the type of crop grown (cash or food), the level of education, being a woman (gender) and income levels greatly influence women farmers’ access to credit

Another view of Mpuga (2010), he employed a panel data analysis of the Uganda household surveys conducted in 1992/1993 and 1999/2000 to study the accessibility to and the characteristics of demand for rural credit He found that the credit market is highly segmented in Uganda; determinants which strongly affect to demand for credit are the value of household assets owned, the level of education, age, location, occupation and other dwelling characteristics He also spotted that the educated and the young are eager to demand for credit while women are likely ashamed to apply for credit or less amount of credit compared to men Additionally the major credit supplies almost come from informal sources; and households which have collateral to secure loans usually tend towards to demand for credit higher

In the study of Hussain and Khan (2011), they examined the determining factors which impact formal and informal credit demand by cotton growers in Bahawalpur They applied the ordinary least squares model and made use of a survey data set of

563 households in eleven villages of Hasilpur, Bahawalpur district, Pakistan

Some findings were that bigger farm size and higher educated households tend to demand credit from the formal financial institutions Meanwhile, the distance of formal sector, transaction cost, corruption and bribe tend to move the farmer from formal to informal sector

Similarly, Menkhoff and Rungruxsirivorn (2011) analyzed whether financial access is improved by Village funds They used the multinomial logit model and the household survey data set of 2200 households in three provinces in North-East Thailand to investigate There were three exciting findings in their study Firstly, the village funds reach the lower income groups better than formal sectors Secondly, the village funds give loans to typical borrowers of informal than formal sectors Finally, credit constraints are alleviated by the village funds

The participation in the credit market, which is affected by both demand and supply side factors, is a different concept from the credit access It was well described in an article of Shete and Garcia (2011) In order to identify the participation proportion of constrained and unconstrained farmers in the agricultural credit market, and to predict the factors which impact agricultural credit market participation, they took advantages of the bivariate probit model and a primary data set collected from the survey of 210 households in Finoteselam town, North Western Ethiopia Based on their study, there were a large proportion of smallholder farmers who were rationed out due to lacking of access Determinants pushing households participate in rural credit markets were household labor force, land size, high dependency ration, contingent expenses and participation in non-farm activities

+ Vietnamese empirical studies

According to Pham and Izumida (2002), they used cross sectional household survey data and the Tobit model to examine the participation of Vietnamese rural household in the rural credit market, and the lending behavior of formal lender in response to the credit demand of households; they also employed the probit model

to investigate the impact of credit on household production Data were collected from the survey of 300 households and conducted in three representatively provinces of Vietnam: Ninh Binh (North), Quang Ngai (Center), and An Giang (South) The results showed that loans from the formal financial institutions are usually for production cultivation, livestock, etc In the other, consumption seems to

be the essential force behind the loans from informal sources Moreover, borrowing from informal sources could be partially for production; and borrowing from formal institutions is an additional Some findings were that accessibility to the formal financial institutions is decisively determined by total farming area and total value

of livestock; additionally it seems that the factors leading to the decision to borrow from informal sources are the dependency ratio of households and total farming area; total farming areas, deposit amount, liquidity position, and number of dependent are the determinant factors of household’s credit status; households, which have high dependence ratio, require more loans to finance their production as well as their consumption

In the other study, Barslund and Tarp (2008) carried out a survey of 932 rural households in four provinces of Vietnam comprising Long An, Quang Nam, Ha Tay and Phu Tho; and combined with information from the Vietnam Household Living Standard Survey 2002 By making use of the two-stage model and the Heckman model, they analyzed the determinants of credit demand and credit rationing for rural households respectively; and found that rural households either who are older

or better educated will access less to credit in the informal sector, whereas households which have a high dependency ratio and bad credit history access towards to more informal credit demand

The following table summarizes the empirical studies above in a more intuitive way

Table 2.1: Early empirical studies on rural credit

1 Pham &

Izumida (2002)

The Tobit model

Data were obtained from the survey of 300 households and conducted in three representatively provinces of Vietnam: Ninh Binh (North), Quang Ngai (Center), and An Giang (South)

The accessibility to the formal financial institutions is

decisively determined by total farming area and total value of livestock; additionally it seems that the factors leading to the decision to borrow from informal sources are the dependency ratio of households and total farming area; total farming areas, deposit amount, liquidity position, and number of dependent are the determinant factors of household’s credit status; households, which have high dependence ratio, require more loans to finance their production as well as their consumption

2 Okurut et al

(2005)

The Heckman two-stage model

Data of 10692 households were surveyed in the Uganda National Household Survey 1999/2000

Determinants which significantly and positively impact informal credit demand are age, dependency ratio, education level, household expenditure, gender and regional location On the other

No Authors Methodology Data Results

hand, factors consisting of age, asset values, regional location and gender are statistically significant and negative impact in term of the supply side

3 Barslund &

Tarp (2008)

The two-stage model (credit demand) and the Heckman model (credit rationing)

ILSSA 2003, VHLSS 2002 Data were collected from 932 rural households in

4 provinces comprising Long

An, Quang Nam,

Ha Tay and Phu Tho

Rural households either who are older or better educated will access less to credit in the informal sector, whereas households which have a high dependency ratio and bad credit history access towards

to more informal credit demand Households seem to

be credit rationed if they have

a bad credit history

4 Guirkinger

(2008)

The logit model with fixed effects

& random effects

A panel data was surveyed in two periods in the rural coastal area of Piura, Peru’s north coast First, there were 547

households surveyed in 1997

Second, 499 households were reinterviewed in

2003

Transaction costs in the formal sector are higher than in the informal sector, it could lead

to the informal effective cost

of a loan cheaper; hence, it could be affordable for some households access to informal loan in term of effective cost

In addition, informal lenders have a more efficient

mechanism in direct screening, monitoring and loan recovery techniques

No Authors Methodology Data Results

compared to formal lenders They can give loans to borrowers with more flexible

& convenient conditions Thus risk adverse households tend

to borrow from informal sources

5 Akudugu et al

(2009)

The logit model

Data were obtained from the survey of 200 women farmers in the Upper East region of Ghana

Rural formal credit has been in favor of women with the proportion of 44 percent compared to men is 56 percent Membership to economic associations, the type of crop grown (cash or food), the level of education, being a woman (gender) and income levels greatly

influence women farmers’ access to credit

6 Mpuga (2010) The probit,

multinomial logit and tobit models

A panel survey data set was used from the Uganda household surveys which was

conducted by UBOS in 1992/1993 and 1999/2000

The credit market is highly segmented in Uganda;

determinants which strongly affect to demand for credit are the value of household assets owned, the level of education, age, location, occupation and other dwelling characteristics The educated & the young are eager to demand for credit while women are likely

No Authors Methodology Data Results

ashamed to apply for credit or less amount of credit

compared to men In addition, the major credit supplies almost come from informal sources; and households which have collateral to secure loans usually tend towards to demand for credit higher

7 Hussain &

Khan (2011)

The OLS model

A survey data set

of 563 households

in eleven villages

of Hasilpur, Bahawalpur district, Pakistan

The bigger household size and educated households tend to demand credit from the formal financial institutions The distance of formal sector, transaction cost, corruption and bribe tend to move the farmer from formal to informal sector

8 Menkhoff &

Rungruxsirivorn

(2011)

The multinomial logit model

A household survey data set of

2200 households

in three provinces

in North-East Thailand

Firstly, the village funds reach the lower income groups better than formal sectors Secondly, the village funds give loans to typical borrowers

of informal than formal sectors Finally, credit constraints are alleviated by the village funds

9 Shete & Garcia

(2011)

The bivariate probit model

A primary data set were collected the survey of 210

There are a large proportion of smallholder farmers who are rationed out due to lacking of

No Authors Methodology Data Results

households in Finoteselam town, North Western Ethiopia

access Determinants pushing households participate in rural credit markets are household labor force, land size, high dependency ration, contingent expenses and participation in non-farm activities

2.4 Conceptual framework

Factors influencing on informal rural credit are grouped into the following grouped factors: (i) a grouped factor representing for demography characteristics includes age, education, gender and occupation, (ii) a grouped factor representing for productive asset consists of age and education, and (iii) a grouped factor representing for productive capital comprises land size, labor force, asset, expenditure and expense on livestock

In order to test hypotheses based upon the relationship between an explanatory variable and independent variables The explanatory variable is considered as probability of farm household access to informal credit

There are two criteria to evaluate credit accessibility of individual: (i) probability of an applicant is selected; (ii) total amount of loan is granted by lender

Firstly, to evaluate the probability of a selected applicant, the explanatory variable is dichotomous showing probability of whether a borrower is selected to lend The first equation is formulated as follows:

The explanatory variable of 𝐴𝑐𝑐𝑒𝑠𝑠𝐻𝐻𝑖 obtains two values:

𝐴𝑐𝑐𝑒𝑠𝑠𝐻𝐻𝑖=1, the ith household is selected to lend

𝐴𝑐𝑐𝑒𝑠𝑠𝐻𝐻𝑖=0, otherwise

The Xi vector includes asset, land size, age, education, dependent, expenditure, livestock, gender, occupation

Secondly, credit accessibility of household is measured by amounts of loan received from lenders In order to evaluate individual’s loan amount, the second equation is formulated as follows

The specific hypothesis for each factor is as follows:

a) Size of land (lnlandsz) (Pham and Izumida, 2002):

It is the total cultivated area in the last 12 months It includes rice cultivation, annual crops, perennial plants and fruit trees If households have more land for cultivation, they will get access to informal sources easier since with high value of harvest they can have much money for repayment Thus, the total cultivated area has positive impact on probability of access to informal credit markets

b) Age of household head (lnage) (Barslund & Tarp, 2008):

This factor could be a negative or positive coefficient It would be a negative sign if household was an old person since older people are more experience, reputation and savings, Lenders could lend money for them due to those advantages Otherwise younger households were mostly risk lovers; they were motivative and eager to catch up new technologies to apply for their production; yet they lack of collateral and saving, lenders could refuse to lend them It is an ambiguous coefficient

c) Education of household head (edu) (Barslund & Tarp, 2008):

If households have higher education, they will gain more knowledge to develop their production In other words, they have more advantages to access to informal financial sector It is therefore, hypothetically expected to have positive coefficient d) Gender of household head (gender) (Akudugu et al, 2009):

This is a dummy variable The value gets one for male and zero for female Gender plays a crucial role in the borrowing process If the household head is female, she is usually assigned to light works and then need a small credit It is expected to have negative sign Reversely, male belongs to hard working types; he has more social relationship than female Hence he has more chance to earn more money and repay the loan The coefficient is expectedly positive

e) Expenditure of household (lnexp) (Shete & Garcia, 2011):

It is expenditure on consumption and non-production, which were aggregated from food, foodstuff, self-produced products, education, healthcare, contingencies, commodities, energies and water supply and sewage system If households expend more, they will need more credit for smoothing consumption and contingencies as well This factor has positive impact on credit access The measure is VND thousand

f) Adult (Menkhoff & Rungruxsirivorn, 2011):

Adults are people whose age is between 15 and 64 years old If a household has many adults, it means that they have a higher probability to repay the loan and then easier to access to the informal sector The coefficient is expectedly positive

g) Dependent (dep) (Menkhoff & Rungruxsirivorn, 2011):

This factor is expected to have negative coefficient, meaning that the larger number of dependents the higher risk of repayment Dependents are people whose age are less than 15 years or above 64 years and not considered to contribute to the labor force of the family

h) Total value of asset (lnasset) (Okurut et al, 2005):

It is total value of fixed and durable asset of household If they have higher value of assets, they have more chance to get loan Thus, it impacts positively on credit demand It is measured by VND thousand

i) Livestock (lnlstk) (Barslund & Tarp, 2008):

Expense on livestock is considered as proxy of livestock variable The more expense on livestock, the higher probability of access to the informal credit market; thus, this coefficient is expected to have positive sign It is measured by VND thousand

j) Farming (farming):

This is a dummy variable It equals to one if occupation of household head is farming and zero if otherwise It is expected to have a negative relationship with credit access since it produces fewer returns than livestock and seemingly depend upon the nature

k) Land used for farming (farmland)( Menkhoff & Rungruxsirivorn, 2011): Expectedly, this coefficient has a negative sign because the more use of land for farming, the less credit-worthiness for borrowers

l) Land used for livestock farming (farmliv) (Barslund & Tarp, 2008):

In the contrary of farmland, this factor is expected to have a positive sign since livestock farming could produce more returns than cultivation

2.5 Chapter summary

This chapter has described a general picture of informal rural credit markets and instruments used in lending activities There were a variety of models which were employed to detect determinants most influence household access to credit A high proportion of informal loan used to smooth consumption and non-production purposes Undeniably farm households have a higher dependency ratio, they have a tendency to require more loans to finance their consumption and contingencies as well Beside that household and farming charateristics, such as age, education,

gender, farming area and expenditure are also significantly influence access to credit in the informal financial sector

CHAPTER 3: DATA AND RESEARCH METHODOLOGY

Literature review and conceptual framework of rural credit have been reviewed in the previous chapter The Vietnamese financial system has experienced with formal sector, semi-formal sector and informal sector Each sector reveals its contribution to different aspects of the picture of rural credit in Vietnam This chapter presents an overview of rural credit market in Vietnam, data used for this study and research methodology

3.1 Background of rural credit markets in Vietnam

Before Doi Moi in 1986 in Vietnam, the State Bank of Vietnam (SBV) was the solitary credit provider This bank played a role as either a central bank or a commercial bank In this period, with the association of the SBV, credit cooperatives intensively handled the credit needs of individuals, cooperatives and communes Rural credit needs existed but at the cooperative level rather than the household level since a cooperative was the smallest unit of production and management

After Doi Moi, the household was introduced as an unit of agricultural production; and economic reforms policies were implemented to spur developement

of social economics in general and financial sector in particular The household credit demand has risen sharply In order to catch up the movement of commercial and financial activities, the SBV was divided into two state owned commercial banks: (i) the Agricultural Development Bank, and (ii) the Commercial Bank of Vietnam

However, these banks served only state owned enterprises; in addition, the credit needs of farmers and rural households were still neglected This result led towards a growth of credit cooperatives by offering high rates of interest on savings Due to the unsuccesful of the collective production system, the collapsed of credit cooperatives and banking system were took place in 1990 - 1991 The Agricultural

Development Bank changed to the Vietnam Bank for Agriculture and Rural Development (VBARD) and worked mainly as a credit provider for rural households

In 1995, in keeping with the program of Hunger Eradication and Poverty Reduction, the Bank for the Poor (VBP) was established as a new complementary funding institution from the VBARD The objective of this bank was to provide the rural poor with the interest credit subsidies In 2003, the VBP separated from the VBARD and nominated as the Vietnam Bank for Social Policies (VBSP) (Takashi, 2009)

Beside VBARD and VBSP, People’s Credit Funds network was mentioned

as a third financial institution in 1993 Three of these financial institutions are essential formal credit suppliers in Vietnamese rural credit markets and concentrating on asset accumulation and loans with production purposes (Barslund

& Tarp, 2008) The amount of loan is used in the formal sector is as the figure 3.1 The formal financial sources are the main credit providers for rural credit markets in Vietnam

Figure 3.1: Rural loan in the formal sector

Source: Author’s calculation in VHLSS 2004, 2006, 2008 & 2010

5.000 10.000 15.000 20.000 25.000 30.000 35.000 40.000 45.000 50.000

Next to the formal sector, there exists microfinance organizations, such as the Vietnamese Women's Union (VWU) and the Vietnamese Farmers' Association (VFA), which offering the rural poor, ethnic minorities and women access to fundamental finance services including savings and loans These organizations were called the semi-formal sector (Takashi, 2009) (APEC, 2011) The distribution of loan amount in the semi-formal sector is as the figure 3.2 The semi-formal sector essentially concentrates on saving mobilization Its mechanism often operates in rural areas and provides financial services with short-term and small amount of loans

Figure 3.2: Rural loan in the semi-formal sector

Source: Author’s calculation in VHLSS 2004, 2006, 2008 & 2010

The informal sector has appeared to contribute to the complete picture of the rural credit market in Vietnam This financial source comes from Group Lending, Moneylenders, Rotating Savings and Credits Associations (ROSCAs), relatives and friends (Pham & Izumida, 2002) Under central planning, the informal financial markets in Vietnam has been suppressed, information about them are very scant The Financial liberalization, which taking place in 1989, contributed to a revival and growth of the informal financial markets Recent studies have revealed the fact that the informal source provides a large number of financial intermediaries In the

500 1.000 1.500 2.000 2.500 3.000 3.500 4.000 4.500

Vietnam context, there are four types of major informal financial participation (Tran, 1998), (Pham & Izumida, 2002) and (Takashi, 2009): (i) Mutual finance is among family members, relatives, neighbors and friends This kind of credit usually

is small amount of money and using for contingencies rarely for buying input

material (ii) Self-help group is spontaneously established by members to help each

other in production and life It operates under the form of ROSCA The other type

of self-help group is collaborates with local formal institutions to reduce transaction costs and enhance peer monitoring loan among groups (iii) Moneylenders are characterized as monopoly lenders because of the absence of an effective formal financial system Experience shows that where access to formal financial sector is limited, the loans are in shortage, they can charge exorbitant interest rates The size

of loan is usually small amount; procedures are simple and convenient; no collateral

is needed and they often operate locally And (iv) traders often use the form of advancing inputs They advance inputs and receive the products with prefixed negotiated rate at harvest In some communes, cooperatives play the role of such traders The informal loan taking part in the informal credit markets is as the figure 3.3

Figure 3.3: Rural loan in the informal sector

Source: Author’s calculation in VHLSS 2004, 2006, 2008 & 2010

2.000 4.000 6.000 8.000 10.000 12.000 14.000 16.000 18.000

The informal credit sector still play a essential role in rural credit markets since (i) the rural poor cannot be affordable for market interest rates, (ii) commercial lenders usually exclude demand of small loans due to high transaction cost, and (iii) collateral is to signal creditworthiness of borrowers to commercial lenders in term of asymmetric information

3.2 Data

All data of the thesis were acquired from the Vietnam Household Living Standard Survey which was carried out in the year 2008, conducted by the General Statistics Office of Vietnam (GSO), and also got helps on technical support of the World Bank (WB) The data for the study only obtained rural population, with total

of 4837 households including 1103 households have credit access and 3734 households do not access to credit, which is representative for the whole country and rural population presented as figure 3.4

The information of the survey was collected through household questionnaires Household information consists of basic demography, education, number of dependents, labor force participation, expense on livestock, expenditure

on consumption and non-production, fixed assets and durable goods, cultivation areas, participation of households in informal credit markets, and especially the information of informal loans had been borrowed during the last 12 months before the interview took place

Data on cultivation areas, expenditure on consumption and non-production used a lot of small questions in details Cultivation areas were summed up from areas of rice cultivation, annual crops, perennial plants and fruit trees Expenditure

on consumption and non-production were aggregated from food, foodstuff, produced products, education, healthcare, contingencies, commodities, energies and water supply and sewage system

self-Figure 3.4: Sample size

Source: Author’s calculation in VHLSS 2008

3.3 Research methodology

3.3.1 Descriptive analysis

In this section, the analysis consists of calculation and comparison of rural credit markets, borrower characteristics, loans and terms of loan by using descriptive statistics The descriptive statistics analysis also provides some insight

of various factors relevant to the use of loans, which will be beneficial to analyses and estimations of the econometric model discussed in the following section

3.3.2 Econometric model

Farm household characteristics of credit access are analyzed in this section Due to using non-randomly selected samples in the VHLSS data set to examine loan amounts of households, the result could be biased estimates by using the OLS analysis because of a missing data problem The Heckman two-stage model (1979), which is to correct the error of sample selection, is used to test hypotheses about the relationship between the dependent and independent variables

The conventional sample selection model (Heckman, 1979) can be written as the form:

1103

3734

Access to creditNot access to credit

𝑧𝑖 = 𝛾𝑤𝑖+ 𝜀𝑖 i=1 N

Where

𝑧𝑖: Informal loan amount

𝑤𝑖: Observed variables relating to the ith characteristics of loan amount, 𝑧𝑖 is observed only for households who received loans already from the informal sector 𝑎: The minimum criterion for getting a loan If total characteristics of borrowers (𝑥𝑖) are below the condition, those people cannot borrow money from informal lenders

𝑑𝑖: 1 - access to credit, 0 - cannot access to credit

𝛾 𝑎𝑛𝑑 𝛽 : Unknown parameter vectors

𝜀𝑖 𝑎𝑛𝑑 𝑢𝑖 : Zero mean error terms

With the assumption that 𝜀𝑖 𝑎𝑛𝑑 𝑢𝑖 follow a bivariate normal distribution and then applied a two-stage estimation procedure purposed by Heckman (1979)

The sample selection model in (1) comprises two equations; the outcome equation representing the desired relationship in the population, the selection equation is to take into account of the non-representative nature of the non-random sample According to the literature (Heckman, 1979), in order to facilitate the identification purpose, the vector 𝑥𝑖 should contain at least one variable which does not appear in vector 𝑤𝑖 The outcome equation describes the relationship between

an outcome in interest 𝑧𝑖∗ and a vector of covariates 𝑤𝑖 and the selection equation describing the relationship between a binary participation decision 𝑑𝑖∗ and another vector of covariates 𝑥𝑖

In order to estimate 𝛾 𝑎𝑛𝑑 𝛽 parameters in the model (1), there are two approaches widely used to estimate the sample selection models under this

and the Heckman two-step procedure

The method frequently used for the sample selection model and also is the best alternative to maximum likelihood is the two-step procedure introduced by Heckman (1979) (Lola et al., 2009), (Cameron & Trivedi, 2009) In the Heckman two-step model, the first step is to estimate the binary selection equation through probit over the whole sample in order to obtain estimates of 𝛽̂ Recall model (1), considers the bivariate normal distribution for the error terms which implying independence of the errors and regressors