PhD dissertation in economics microcredit and welfare of the rural households in vietnam

Objective 4: Role of microcredit in mitigating the effects of health shocks 145 8.4.1.. Since its birth, microcredit programs provide the poor with opportunities to be more accessible to

MINISTRY OF EDUCATION AND TRAINING UNIVERSITY OF ECONOMICS HO CHI MINH CITY

*****

Pham Tien Thanh

MICROCREDIT AND WELFARE OF THE RURAL HOUSEHOLDS IN VIETNAM

PHD DISSERTATION IN ECONOMICS

HO CHI MINH CITY, 2018

MINISTRY OF EDUCATION AND TRAINING UNIVERSITY OF ECONOMICS HO CHI MINH CITY

*****

Pham Tien Thanh

MICROCREDIT AND WELFARE OF THE RURAL HOUSEHOLDS IN VIETNAM

Major: Development Economics Code: 9310105

PHD DISSERTATION IN ECONOMICS

ACADEMIC ADVISORS

1 Dr PHAM KHANH NAM

2 Assoc Prof Dr NGUYEN HUU DUNG

HO CHI MINH CITY, 2018

COMMITMENTS

I hereby declare that this dissertation is my own work and it has not been previously submitted for a degree elsewhere While my work contains information from works done by other authors, all these references have been explicitly cited

To the best of my knowledge, I would like to certify that the above statements are true

The dissertation has been done under the supervision of Dr Pham Khanh Nam and Assoc.Prof Dr Nguyen Huu Dung, the University of Economics, Hochiminh City This work was also done under the advice of Assoc Prof Dr Katsuhiro Saito, the University of Tokyo during my fellowship period granted by the World Bank Robert S McNamara at the University of Tokyo, Japan

Pham Tien Thanh

ACKNOWLEDGEMENT

I would like to express my deepest gratitude to my academic supervisors, reviewers, colleagues, family and friends for their support during the time I conduct my Ph.D dissertation

TABLE OF CONTENTS

COMMITMENTS i

ACKNOWLEDGEMENT ii

TABLE OF CONTENTS iii

ABBREVIATIONS viii

LIST OF TABLES ix

LIST OF FIGURES xi

CHAPTER 1 - INTRODUCTION 1

1.1 Backgrounds 1

1.2 Research problems 3

1.2.1 Accessibility to microcredit 3

1.2.2 Microcredit and adoption of improved rice varieties 4

1.2.3 Impact of microcredit on households’ welfare 6

1.2.4 Microcredit and health shock 8

1.3 Research objectives 9

1.4 Scopes of research 9

1.5 Research data 10

1.6 Research methods 10

1.6.1 Quantitative methods 10

1.6.2 Qualitative methods 11

1.7 Contribution of the research 11

1.7.1 Theoretical contributions 11

1.7.2 Practical contributions 12

1.8 Organization of the research 13

CHAPTER 2 – LITERATURE REVIEW 14

2.1 Overview of microcredit 14

2.1.1 Some definitions 14

2.2.2 Characteristics of microcredit 15

2.2 Theory 16

2.2.1 Overview of rural credit market 16

2.2.2 Asymmetric information 19

2.2.3 Theories on accessibility to credit 21

2.2.3.1 Theory of Choice (Demand for Credit) 22

2.2.3.2 Theory of Credit Rationing 23

2.2.4 Theoretical framework of outreach of microcredit 26

2.2.5 Farm household model under credit constraint 27

2.2.5.1 Producer Problem 28

2.2.5.2 Consumer Problem 28

2.2.5.3 Worker Problem 29

2.2.5.4 Farm Household Model with Credit 30

2.3 Empirical studies 31

2.3.1 Determinants of rural households’ accessibility to microcredit 31

2.3.2 Impact of microcredit on adoption of the improved rice varieties 36

2.3.3 Impact of microcredit on living standard 38

2.3.1 Empirical studies in the world 39

2.3.2 Empirical studies in Vietnam 40

2.3.4 Impact of health shocks and role of microcredit 43

2.3.4.1 Impacts of health shocks 43

2.3.4.2 Strategies to cope with Health Shocks and Role of Microcredit 44

2.4 Definitions of welfare 47

2.5 Logical framework of impact evaluation of microcredit on welfare 48

CHAPTER 3 – OVERVIEW OF MICROCREDIT 51

3.1 Microcredit in the world 51

3.2 Microcredit in rural Vietnam 53

3.2.1 Microcredit providers 53

3.2.2 Credit activities and clients 58

3.2.3 Poverty reduction role of microcredit 60

CHAPTER 4 – DETERMINANTS OF ACCESSIBILITY TO MICROCREDIT 61

4.1 Research method 61

4.1.1 Estimation strategy 62

4.1.2 Selection of variables 65

4.1.2.1 Dependent variable 65

4.1.2.2 Explanatory Variables 65

4.1.3 Research hypotheses 67

4.2 Research data 69

4.2.1 Data source 69

4.2.2 Research sample 70

4.2.3 Descriptive statistics 70

4.3 Results and discussions 73

CHAPTER 5 – MICROCREDIT AND ADOPTION OF IMPROVED RICE VARIETIES 81

5.1 Research method 81

5.1.1 Estimation strategy 81

5.1.2 Selection of variables 85

5.1.2.1 Dependent variables 85

5.1.2.2 Independent variables 86

5.1.3 Research hypotheses 87

5.2 Research data 87

5.2.1 Data source 87

5.2.2 Research sample 87

5.2.3 Descriptive statistics 88

5.3 Results and discussions 90

CHAPTER 6 – IMPACT EVALUATION OF MICROCREDIT ON WELFARE 102 6.1 Research method 102

6.1.1 Estimation strategy 102

6.1.1.1 Propensity Score Matching (PSM) 102

6.1.1.2 Difference in Difference (DID) 103

6.1.1.3 PSM-DID 104

6.1.2 Selection of variables 105

6.1.2.1 Microcredit variable 105

6.1.2.2 Variables for calculation of propensity score 106

6.1.2.3 Welfare variables 106

6.1.3 Research hypotheses 107

6.2 Research data 108

6.2.1 Data source 108

6.2.2 Research sample 108

6.2.3 Descriptive statistics 109

6.3 Results and discussions 109

6.3.1 Propensity score and balancing test 109

6.3.2 Average impact of microcredit program 111

CHAPTER 7 – ROLE OF MICROCREDIT IN MITIGATING THE EFFECTS OF HEALTH SHOCKS 118

7.1 Research method 118

7.1.1 Estimation strategy 118

7.1.2 Selection of variable 120

7.1.2.1 Dependent variables 120

7.1.2.2 Health shocks variables 120

7.1.2.3 Microcredit variable 121

7.1.2.4 Other control (explanatory) variables 122

7.1.3 Research hypotheses 123

7.2 Research data 123

7.2.1 Data source 123

7.2.2 Research sample 124

7.2.3 Descriptive statistics 124

7.3 Results and discussions 126

7.3.1 Are health shocks unanticipated? 126

7.3.2 First stage results using instrumental variables 127

7.3.3 Illness/ Disease/ Injury of working-age members (HS1W) 130

7.3.4 Illness/ Disease/ Injury of any members (HS1A) 133

CHAPTER 8 – CONCLUSIONS 135

8.1 Objective 1: Determinants of accessibility o microcredit 135

8.1.1 Summary 135

8.1.2 Policy implications 136

8.1.3 Limitations and further studies 138

8.2 Objective 2: Microcredit and adoption of improved varieties 139

8.1.1 Summary 139

8.2.2 Policy implications 140

8.2.3 Limitations and further studies 141

8.3 Objective 3: Impact evaluation of microcredit on welfare 142

8.3.1 Summary 142

8.3.2 Policy implications 143

8.3.3 Limitations and further studies 144

8.4 Objective 4: Role of microcredit in mitigating the effects of health shocks 145 8.4.1 Summary 145

8.4.2 Policy implications 146

8.4.3 Limitations and further studies 146

REFERENCE 147

Reference in the word 147

Reference in Vietnam 165

Websites 169

APPENDIX 171

ABBREVIATIONS

LIST OF TABLES

Table 2.1 - Welfare indicators 47

Table 3.1 - Microcredit trend in the world 52

Table 3.2 - Top 10 microcredit providers in the world 52

Table 3.3 - Microcredit providers in Vietnam 59

Table 3.4 - Living standards of clients before and after borrowing 60

Table 4.1 - Variables for analysis of determinants of accessibility to microcredit 67

Table 4.2 – Description of loans obtained 71

Table 4.3 - Characteristics of variables for analysis of determinants of accessibility to microcredit 72

Table 4.4 - Estimated results of determinants of accessibility to microcredit – Whole sample 75

Table 4.5 - Estimated results of determinants of accessibility to microcredit – Non-poor sample 79

Table 4.6 - Estimated results of determinants of accessibility to microcredit – Poor sample 80

Table 5.1 - Variables used for analyzing impact of microcredit on adoption 86

Table 5.2 - Characteristics of adoption variables 88

Table 5.3 - Characteristics of variables that affect adopting decision 89

Table 5.4 - Microcredit and adoption of improved rice varieties - IV 2SLS using 3 IVs 91

Table 5.5 - Microcredit and adoption of improved rice varieties - IV 2SLS using 1 IV 91

Table 5.6 - Microcredit and adoption of improved rice varieties 93

Table 5.7 - Microcredit and adoption of improved rice varieties – Poor and non-poor sample 101

Table 5.8 - Impact of microcredit on other investment activities 101

Table 6.1 - Definition of welfare variables 107

Table 6.2 - Probit estimations for calculation of propensity score 110

Table 6.3 – Balancing test 111

Table 6.4 - Impact of microcredit on income and total output value 112

Table 6.5 - Impact of microcredit on food expenditure 116

Table 6.6 - Impact of microcredit on accumulation of durable assets 117

Table 7.1 - Types of health shock 121

Table 7.2 - Microcredit borrowed by households 122

Table 7.3 - Descriptive statistics of variables for used analysis of health shocks 125

Table 7.4 - Unanticipation of health shocks 127

Table 7.5 – First stage results using IV for HS1W 128

Table 7.6 – First stage results using IV for HS1A 128

Table 7.7 - Impact of health shocks on income and labor and the role of microcredit using VFE 129

Table 7.8 - Impact of health shocks on consumption and the role of microcredit using VFE 129

LIST OF FIGURES

Figure 2.1 – Credit constraint and unconstraint 25

Figure 2.2 - Analytical framework of impact of microcredit on welfare 42

Figure 2.3 - Conceptual framework of impacts of health shock 44

Figure 2.4 - Logical framework on impact evaluation of credit on welfare 49

Figure 2.5 – Analytical framework on accessibility to microcredit and its impact on welfare 50

Figure 3.1 – Microfinance providers in Vietnam 53

CHAPTER 1 - INTRODUCTION

1.1 Backgrounds

Poverty reduction, access to education, clean water, sanitary, health care, etc are the top concerned among 17 Sustainable Development Goals - SDGs) which formerly was Millennium Development Goals (MGDs) They are the important issues and attract the attention of every country, especially the developing countries like Vietnam During the past decades, Vietnam has obtained remarkable achievements in the socio-economic development and poverty reduction

Statistics show that the poverty rate has declined from 37.4 percent in 1998

to 5.8% in 2016 (Demombynes & Vu, 2015; GSO, 2016) The report also shows that the poverty rate in 2016 in rural areas (7.5%) is nearly four times higher than that in urban areas (2%)1 The statistics indicate that a large number of rural households still live in poverty under poor living standards and suffer from the lack

of socio-economic opportunities Income of the households in rural area is rather low In particular, a report by GSO (2016) shows that per capita income per month

of the rural households is, on average, 2,437 thousand VND In the rural, per capita income per month of the bottom quintile (the poorest) is 676 thousand VND, the second bottom (the near-poor) 1,246 thousand VND and the top quintile (the richest) 5,669 thousand VND Therefore, it indicates that income of the poorest is around 8.4 times lower than that of the richest, which is a great gap Consumption

of the rural households is quite low In particular, a report by GSO (2014) shows that per capita expenditure per month of the rural households is, on average, 1,609 thousand VND More specific, per capita expenditure per month of the bottom quintile (the poorest) is 834 thousand VND, the second bottom (the near-poor) 1,154 thousand VND and the top quintile (the richest) 2,576 thousand VND, which indicates a big gap between the poorest and the richest in terms of expenditure

1 New poverty line (Measured by income per capita per month) in 2016 is 630 and 780 thousand

VND in rural and urban areas, respectively

Poverty reduction, income improvement and good health, especially in the rural area, are the most concerned issues among the 17 Sustainable Development Goals In the world and in Vietnam, many programs and activities have been implemented to achieve these goals, for instance, training, agricultural extension, agricultural technology, finance, job creation, etc Among them, credit is considered

as the most important factors in income-generating activities since the rural households, especially the poor, normally face financial constraint Without accessibility to formal credit support, poor farmers are unable to expand their production and improve their living conditions (Li et al., 2011)

As above-mention, a large number of rural households in Vietnam are still living in poverty The poor have high demand for credit; however, credit market in rural Vietnam normally faces the problem of demand excessing supply Thenceforth, the poor have difficulty in accessing formal credit due to the lack of collateral In rural Vietnam, formal credit is mainly provided by the Vietnam Bank for Agriculture and Rural Development and Vietnam Bank for Social Policies (Ho

& Duc, 2015), and recently by microfinance institutions

Inability to obtain formal credit has inhibited farmers from improving their living conditions Therefore, farmers may rely on informal credit sources to support their production and consumption even though they must pay a much higher interest rate, which may push them deeper into ‘vicious debt circle’ with less repayment capacity The reasons why farmers still accept the high interest rate from informal sources are its collateral-free merit and quick borrowing procedure

Since its birth, microcredit programs provide the poor with opportunities to

be more accessible to formal and semi-formal credit sources (Li et al., 2011) Some merits of microcredit are collateral-free and to serve the poor By definition, microcredit is defined as a small loan granted to the poor so that they can run production or do business to generate income and improve their well-being (Microcredit Summit, 1997) Microcredit is found by many researchers and practitioners to an effective instrument to improve well-being of the poor

Therefore, this research aims to examine accessibility to microcredit and its economic impacts on welfare of rural households Thenceforth, relevant policies will be suggested to support the poor with better access to microcredit sources and improve the effectiveness of microcredit in investment, thereby increasing their living standard Therefore, to capture these primary research objectives, the following specific objectives will be included and analyzed:

1.2 Research problems

1.2.1 Accessibility to microcredit

Morduch and Haley (2002) state that credits can help the poor to improve their living standards or at least cover their living expenses However, a research by Brau and Woller (2004) find that the poor have difficulties in accessing to formal credit sources; particularly, the poor in the developing countries have more difficulties in accessing than those in the developed countries In Vietnam, many rural households normally have difficulties in accessing to credit, especially poor households, households in remote areas, the minority ethnic groups, or households operating in such fields with high risk as aquaculture, etc., These households always have high demand for credit (Duong & Izumida, 2002) but they have some difficulties in borrowing from semi-formal and formal credit sources such as banks

or financial institutions Thenceforth, many households have to borrow from informal credit sources such as friends, relatives, money-lender, etc The Government has taken a lot of effort to support the rural households with accessing

to semi-formal and formal credit but the result is still rather limited To increase efficiency of financial services as well as better serve the poor, it is essential to understand the factors that affect participants in microcredit of the rural households, thereby implying appropriate policies Therefore, this study is conducted to investigate the factors that affect the accessibility to microcredit sources

The objective of microcredit programs is to serve the poor However, a research by Nguyen (2008) finds that the number of poor households with access to

microcredit program is lower than that of the non-poor Thenceforth, this research will investigate whether the poor households are better able to participate in microcredit programs that the non-poor

In addition, one of the attributes of microcredit programs is that the target clients are women Therefore, this research also investigates the role of gender in participation in microcredit In particular, this research examines whether the female household heads can have better accessibility to microcredit sources than the male counterparts This objective is conducted using models with binary dependent variable such as Probit or Logit

Research question 1: What are determinants of accessibility to microcredit

of rural households? Do poverty status and gender matter?

1.2.2 Microcredit and adoption of improved rice varieties

Agriculture plays a crucial role in the economy of every country, especially the developing countries In Vietnam, agricultural sector contributes nearly 33 billion USD to gross domestic product (GDP) in 2015, around 17-percent share (World bank, 2016) In addition, nearly half (46.8%) of the labor force in Vietnam are employed in agriculture in 2013 (World bank, 2016) Agriculture contributes to ensure food security and generate income for economic development (Datt & Ravallion, 1996) Moreover, in the developing countries, agriculture is a major income source of rural households that derives from domestic sales and export (Singh et al., 1985) Therefore, the improvement of quantity and quality of agricultural output is a focal point of the government in the developing countries (Bonnin & Turner, 2012) However, the population in the world as well as in the developing countries has increased more and more and the urbanization has been so rapid, which results in less land resource for agricultural production Thenceforth, productivity improvement via expanding land area is no longer appropriate Therefore, the most possible solution to enhance production output is to adopt new high-yielding technology

Rice is considered one of the main staples in the world and developing countries like Vietnam In Vietnam, total area of rice farming is around 7,790 thousand hectares with total production output of 43,609 thousand ton in 2016 (GSO, 2016), which makes rice rank the first among annual crop Rice is the leading cultivar in terms of export volume with 4,835 thousand ton and rank second

in terms of export value with 2.1 billion USD (GSO, 2016) Rice production is an important source of export earnings for rice economies like Vietnam Therefore, enhancing quantity, quality and value of rice is the urgent and important issues

Literature documents that adoption of improved rice varieties is considered

as an effective to enhance productivity due to their attributes of short duration, high yield and climate-resistance (Sall et al., 2000) However, the adoption of improved varieties may only become effective in enhancing productivity when being combined with new fertilizers and modern cultivation practices (Karanja et al., 2003)

In practice, adoption of new technology normally incurs high initial cost and high risk, which inhibit the farmers’ decision toward adoption, especially the farmers with financial constraint Several studies find that credit can facilitate the households’ investment in agricultural innovation in general and improved varieties

in particular This can be explained via two mechanisms (Eswaran & Kotwal, 1990; Zeller et al., 1997; Diagne et al., 2000): (1) Credit relaxes financial constraint of the farmers, and thereby they can invest in agriculture; (2) Credit plays a role as coping strategies with shocks, including agricultural shocks such as natural disaster, pestilent insect, etc

In fact, adoption of improved cultivar varieties may be affected by farmers’ perception, education, risk attitude, knowledge and management of agricultural extension center, and soil conditions of the area These models are complicated cultivation practice and require more time, labor and capital than traditional strategies since farmer need to pay more for new and certified agricultural inputs such as seeds or fertilizers (Chi, 2008; Dung & Thanh, 2017)

The previous studies only investigate the determinants of adoption of new varieties, but do not focus on the role of credit on adopting decision Moreover, most of the previous studies observe the behaviors of all farmers while there are very few studies that investigate in detail the behaviors of farmers with and without financial constraints or poor or non-poor status In addition, most of previous studies only investigate the determinants of whether to adopt the improved varieties There are still limited number of studies on the decisions of farmers in the two stages, including whether to adopt and intensity of adoption

In Vietnam, there has been no empirical studies investigating the relationship between microcredit and adoption of improved rice varieties This study examines the role of microcredit in two-stage decision of the farmers on adoption of improved varieties as well as investigates in detail the behaviors of each specific group (poor and non-poor groups) Therefore, instead of a simple model, this research employs two-stage models such as Double-Hurdle (DH) or Heckman to clarify the farmers’ adoption decisions in two stages For robustness check, such models as Tobit are also applied for comparison

Research question 2: Does microcredit enhance the farmers’ decision on

adoption of improved rice varieties in two stages, including whether to adopt and how much to adopt?

1.2.3 Impact of microcredit on households’ welfare

Microcredit programs have been applied in many countries as a tool of poverty reduction and hunger eradication The impact of microcredit on living standard of rural households has been studied by many scholars, practitioners and policymakers However, their findings are mixed and inconsistent

Khandker (1998) and Yunus (2003) consider microcredit as a tool for the goal of fighting poverty and improving welfare Microcredit is also found to have positive and significant effect on welfare of the rural households via improving their consumption and/or consumption (Li et al., 2011b; Mahjabeen, 2008) It is also

proved that households with access to microcredit can improve their children’ health condition of (Pitt et al., 2003) or education (You & Annim, 2014)

However, some researchers argue that microcredit has insignificant effect on households’ living standard Coleman (2006) finds no evidence for the impact of microcredit on less wealthy households’ living standards, while Khandker and Koolwal (2013) contrast that only smaller landholders benefit from microcredit via increasing their agricultural income Takahashi et al (2010) conclude that microcredit has no significant impact on various outcomes, except for sales from self-employment for the non-poor and schooling expense for the poor, thereby indicating that microcredit has no immediate impact on poverty reduction

In Vietnam, there have been many studies about the impact of microcredit on the welfare of the rural households The empirical studies show that microcredit improves households’ welfare via increasing income, expenditure, self-employment profits or contributing to poverty reduction (Nguyen, 2008; Quach, 2017; Lensink

& Pham, 2011; Ho & Duc, 2015) In some cases, microcredit is found to have no impact on income, but only consumption (Phan et al., 2014)

Literature shows that the impact of microcredit on welfare is still in argument and the results are mixed among research and depend on indicators used

to measure welfare While there have been many researches about the role of microcredit in Vietnam, none of these studies investigate its impact on multi-dimensional aspects of welfare Therefore, this research is conducted to examine the role of microcredit on households’ welfare reflected via income, food consumption, and accumulation of assets To capture to better results, this research applies various advanced impact evaluation techniques such as Propensity Score Matching (PSM), Difference in Difference (DID) and PSM-DID

Research question 3: Does microcredit improve rural households’ welfare,

measured by income, total production value, food consumption and asset accumulation?

1.2.4 Microcredit and health shock

A research by Banerjee and Duflo (2012) finds that when health is improved, people can study or work better, which thereby improves their future income In the developing countries, the social security system is not well-organized Thenceforth, when households face any health problem, they normally face big trouble, especially the poor Sytemmatic review by Alam and Mahal (2014) reveals that health problems may have impact on household via some channels For instance, when the workers in the family cannot work due to health problems, it will lead to

an increase medical expense and a decrease the household’s income The same outcome may happened when care-givers cannot work to give cares to other sick/disease member To alleviate such burden, the households with health problems will normally use a certain coping/mitigating strategy However, when the social security system is less developed and the formal coping/mitigating strategies are not availables, the households may take advantage of informal coping strategies such as reduced consumption, informal loan, or selling assets

Regarding the role of microcredit in coping with health shock, there have been few studies (Gertler et al., 2009; Islam & Maitra, 2012) These studies state that households with access to microfinance can smooth their consumption better than those without any coping strategies

In Vietnam, impact of health shock has been researched in few studies (Wagstaff & van Doorslaer, 2003; Wagstaff, 2007; Mitra et al., 2015; Okonogi et al., 2015) However, neither of researches in Vietnam investigates the role of microcredit in mitigating the impact of health shocks, especially in relation with other outcomes Therefore, this research has two objectives The first aim is to examine the impact of health shocks on the households’ change in income, consumption and intra-household labors The second and more important aim is to investigate the role of microcredit in mitigating the consequence after an onset of health shocks Regarding the inter-relation between microcredit and mobility of intra-household labors after an onset of health shocks, there seems to be lack of

literature on this issue, which thereby makes this research become innovative This study employs regression with Village Fixed Effect (VFE) using the data from VHLSS in 2010 and 2012 to make comparison and capture the best results

Research question 4: Does microcredit mitigate the adverse effects of health

shocks on income, consumption and intra-household labor mobility?

1.3 Research objectives

The primary research objective is to investigate the accessibility to microcredit programs and its impacts on welfare However, it is important to capture the impact mechanism of microcredit program Thenceforth, following

logical framework of impact evaluation in Section 2.5, this research includes the

following specific objective:

(1) Investigate the determinants of accessibility to microcredit of the rural households; in particular, whether microcredit targets the poor and women borrowers Further investigation is also conducted on the poor and non-poor sample

(2) Examine the role of microcredit in enhancing farmers’ decision on adoption of improved rice varieties in both stages, including whether to adopt and how much to adopt The study decomposes research sample into the poor and non-poor for further investigation

(3) Evaluate the impact of microcredit on the rural households’ living standards, measured by various indicators such as income, total production value, food consumption, and asset accumulation

(4) Investigate the mitigating role of microcredit when households face any health shocks; in particular, whether microcredit reduces the adverse effects of health shocks on income, consumption and intra-household labor mobility

1.4 Scopes of research

Microcredit is widely provided in urban and rural areas Both banks and Microfinance Institutions (MFIs) provide microcredit In addition, many

indicators can be used to define microcredit, welfare and health shocks Due to availability of data, the research scopes are as follows:

• This research focuses on the households in the rural areas in Vietnam

• Microcredit used for analysis is defined as collateral-free and small loan (less than 100 million VND) granted by the formal credit providers such as banks, PCF and other mass organizations

• This research focuses on the microcredit used for production or doing business Thenceforth, it may capture the long-term and sustainable impact of microcredit on welfare

• Welfare indicators used for analysis include total output value, revenue, income, consumption, accumulation of assets, and non-working-age labor

• Two types of health shocks are used for analysis, including illness/ disease/ injury of any members and working-age members

1.5 Research data

Quantitative data This research applies two datasets for quantitative

analysis, including Vietnam Household Living Standards Survey (VHLSS) in 2010 and 2012, and Vietnam Access to Resources Household Survey (VARHS) in 2012 and 2014 VARHS is applied to investigate the first three objectives while the fourth objective is analyzed using VHLSS

Qualitative data In-depth interview and focus group are conducted in rural

areas in Tra Vinh and Long An provinces to supplement the results from econometric model and to imply policy

(DH), Tobit and Heckman models with cross-sectional data are used to examine how microcredit affects farmers’ decision toward adoption of improved rice

varieties (Research objective 2) Such impact evaluation techniques as Difference

in difference (DID) and PSM-DID with panel data are applied to assess the effects

of microcredit on living standard (Research objective 3) Regression with

Village-Fixed-Effect (VFE) with panel data is employed to analyze the role of microcredit

in mitigating the consequence of health shocks (Research objective 4)

1.6.2 Qualitative methods

According to Merriam (1998), Bogdan and Biklen (1992), and Creswell (2003), qualitative research may be conducted using: observation, In-depth Interview (Face to Face, Focus group, phone), Media (Photo, Recording) This research employs in-depth interview and focus group to capture the deeper understanding about the research problems Qualitative research is conducted on the subjects associated with microcredit programs, including: (1) Borrowers from microcredit provided by VBSP, (2) village Officials, staff of Women Union, staff of Farmer Union, etc at some communes; (4) Specialists in the fields of rural finance

1.7 Contribution of the research

1.7.1 Theoretical contributions

The dissertation is an empirical work, which mainly uses applied econometrics and models from previous studies to testify for the case of Vietnam However, the dissertation still makes some contributions to literature, including:

• The dissertation modifies an analytical framework on mechanisms through which microcredit can affect farmers’ decisions on adoption of improved rice varieties or other investment choices, then impact households’ welfare, and finally mitigate the adverse effect of health shocks

• The dissertation modifies and test empirical models of the determinants of accessibility to microcredit

• The dissertation also makes contributions on literature about the effects of microcredit on farmers’ decision on whether to adopt and how much to adopt improved rice varieties

• The dissertation examines the role of microcredit in mitigating the effects of health shocks in relation with intra-household labor mobility, which then makes a significant contribution to literature in this field

• In addition, for each research objective, different applied econometrics methods are employed for robust checks to capture better results Also, the dissertation decomposes research sample to different groups for further investigation to capture a thorough understanding

• Finally, the dissertation contributes to the academic aspect regarding policy analysis methods using IV 2SLS and PSM-DID

1.7.2 Practical contributions

Poverty alleviation, good education, gender equality and good health are the top concerns in the Sustainable Development Goals (SDG) Microcredit is considered as an effective tool for improving household well-being Therefore:

• The dissertation will be an evidence for the local authorities, policy-makers and practitioners to recognize and understand the role of microcredit in fostering investment and then improving living standard of the rural households Thenceforth, they can disseminate the similar programs to better support the rural households

• Moreover, the dissertation provides policy implications toward improving households’ accessibility to microcredit to increase the outreach of microcredit, especially to the poor or the disadvantaged

• Other relevant supporting policies, but not directly related to microcredit, are also implied to improve the effectiveness of microcredit programs

• The empirical findings from the dissertation will suggest important policy implications related to microcredit programs, not only in Vietnam but also

in other emerging, transition and low- and middle- income economies

1.8 Organization of the research

Chapter 1: This chapter introduces research problems, research methodology, research questions and objectives, and contributions of the research

Chapter 2: This chapter presents literature review, including theoretical and empirical studies associated with to four research objectives, including: Determinants of accessibility to microcredit, (2) Microcredit and adoption of improved rice varieties, (3) Impact evaluation of microcredit on welfare, and (4) Role of microcredit in mitigating the effects of health shocks

Chapter 3: This chapter describes some overviews of microcredit and rural financial market in the world and Vietnam

Chapter 4: This chapter presents the contents related to the first research objective (Determinants of Accessibility to Microcredit), including method, data, results and discussions

Chapter 5: This chapter presents the contents related to the second research objective (Microcredit and Adoption of Improved Rice Varieties), including method, data, results and discussions

Chapter 6: This chapter presents the contents related to the third research objective (Impact Evaluation of Microcredit on Welfare), including method, data, results and discussions

Chapter 7: This chapter presents the contents related to the third research objective (Role of Microcredit in Mitigating the Effects of Health Shocks), including method, data, results and discussions

Chapter 8: This chapter summarizes main findings in the research, gives policy implication, mentions limitations and further research

CHAPTER 2 – LITERATURE REVIEW

This chapter describes literature of microcredit, including definition, characteristics, theory and empirical studies, thenceforth establishing analytical

framework This chapter includes five sections Section 2.1 presents some overviews of microcredit Section 2.2 presents some theories associated with credit and microcredit Section 2.3 describes empirical studies relevant to four research objectives Section 2.4 briefly discussion definition of welfare Section 2.5 present

the logical framework connecting the general and specific research objectives

2.1 Overview of microcredit

2.1.1 Some definitions

Microfinance is considered as an effective tool for fighting poverty and eradicating hunger Microfinance was initiated by the Grameen Bank - founded and managed by Prof Muhammad Yunus, who gained the Nobel Peace Prize in 2006

is defined as microfinance services granted by formal and semi-formal sectors

Morduch (1999): “Microfinance is defined as the provision of small-scale financial services for the poor”

Dasgupta and Rao (2003): “Micro-finance is a financial service of small quantity provided by financial institutions to the poor These services may include credit, savings, insurance, leasing, money transfer, etc that are provided to clients

to meet their financial needs.”

- Microcredit

The Microcredit Summit (1997): “Microcredit is a program which extend

small loans to very poor people for self-employment projects that generate income, allowing them to care for themselves and their families”

Ault and Spicer (2008): “microcredit is a model of lending that give small loans to the poor who lack access to formal financial institutions”

Dash (2012): “Microcredit symbolizes small loans extended to very poor people for self-employment projects that generate income, allowing them to care for themselves and their families”

2.2.2 Characteristics of microcredit

Literature document that microcredit may include the following attributes:

(1) Small loan size The term “small loan" may vary in each country In

Vietnam, amount of formal microcredit ranges from less than 10 up to 100 million VND (Khoi et al, 2013; Thanh, 2017; Thanh & Dung, 2017)

(2) Targeting the poor Microcredit targets the poor and low-income

households (Morduch, 1999; Du, 2004; Dasgupta & Rao, 2003)

(3) No collateral Collateral is normally required to reduce default risk by the

formal lenders However, the poor lack of assets for collateral, which inhibits them from accessing formal credit Microcredit is collateral-free, which thereby increases the poor’s accessibility to credit (Li et al., 2011a; Thanh, 2017)

(4) For investment purpose Borrowers may use microcredit to invest in

either farm or on-farm activities (Khandker & Koolwal, 2016)

(2) High interest rate The rate may vary from 15 to 35 percent per year

However, Li et al (2011a) state that rural households in China prefer microcredit to other credit because of its affordable interest

(6) Group-lending scheme Due to no requirement for collateral, following

this scheme, the members of credit group will monitor the others and be responsible for the latter’ repayment If a member defaults the loan, the remaining members

may have difficulty in obtaining the loan from the program in the future lending scheme is considered as a “social collateral” Thenceforth, the lenders can reduce risks of default by the borrowers and transaction costs (Anderson & Nina, 2000; Besley & Coate, 1995)

Group-(7) Targeting women Around 74 percent of all microcredit borrowers in the

world are women (Cheston & Kuhn, 2002) Thenceforth, it indicates that microcredit programs mostly target female clients

Based on the characteristics of microcredit and the availability of data, in this research, microcredit used for analysis of the first three objectives is defined as small amount (less than 100 million dong) from formal sources, used for the purpose of production or self-employment For the final research objective, the purpose of microcredit is extended, including for consumption

2.2 Theory

2.2.1 Overview of rural credit market

Credit markets are intermediation between savers and borrowers Credit markets are characterized by such problems as risk and information, and they are distinguished from good and services markets (Hoff and Stiglitz, 1990, 1997) Similar to credit market in general, rural credit markets are normally imperfect and more so in the less developing countries (LDCs) Some plausible explanations are:

• The transaction of ‘buying’ (borrowing) and ‘paying up’ (repayment) are

made separately in time Normally, ‘buying’ process is made beforehand; the repayment is made latter (normally, after the investment cycle) That is, many livelihood strategies, especially agricultural production are spread over time, in particular farmers invest today, but reap outcomes tomorrow

• Lenders are unable or have limited ability to monitor what borrowers do with

a loan (e.g use for too risky projects or for wrong purposes) There is a possibility of default because borrowers may be unable to repay due to uncertain outcomes in agricultural production, loss in business or losing jobs

In some cases, if borrower feels better off not returning the loan, he will not The problem of willing to repay occurs if contract enforcement is weak, and taking borrowers to court tends to be more costly Another reason is that the borrower’s liability is limited, When the borrowers default the loan, credit officers may share a part of liability

• Due to imperfect information, inability or limited ability to monitor and

possibility of default, there exists large share of informal credit providers in developing countries This is because informal sector has stronger enforcement and better information and monitoring

In rural credit market, lenders seem to be more careful than borrowers Lenders are, especially informal lenders, use indirect or direct screening mechanisms to address problems of incentives and enforcement To prevent adverse selection or/and moral hazard, the lenders normally screen and monitor the borrowers, obtain collateral from borrowers; or threaten to reduce the future loan

• Borrowers differ in their probability of default, and it is costly to specify the

risk level for each borrower This problem is conventionally known as

screening Lenders screen the borrowers regarding the available information

• It is costly to make sure that borrowers take actions which make higher

probability of repayment This is called incentives problem

• It is difficult to make repayment compulsory This problem is known as

enforcement

Rural credit market in developing countries is characterized by the variation

in interest rate The annual interest rate in rural area of LDCs is normally higher than that of developed countries Informal sector charges higher rate than formal sector; for instance, Siamwalla et al (1990): study of rural credit markets in Thailand, found informal sector annual interest rate to be 60% whereas formal sector rate ranged from 12-14% This may be explained partly by monopoly Informal credit provider can charge different interest rate on different borrowers using price discrimination strategy Monopoly power is established by natural entry

barriers which is reflected via transaction cost However, monopoly power of informal lender is restricted because of the presence of formal sector (for instance, public banks) and recently the semi-formal sectors

Borrowers can borrow only up to a limited amount at a given interest rate, and will not be granted a greater amount even if they are willing to pay higher rate The poor normally face difficulties in borrowing at any interest rate This indicates that credit market does not follow standard demand-supply equilibrium

In credit market, sources of demand for credit include (1) Fixed capital for new startups or expansion of production (machines, buildings); (2) Working capital

for on-going production because of time lag between putting in inputs and reaping

outputs; and (3) Consumption credit for consumption smoothing Among them,

fixed capital is mainly vital for overall growth and working capital and consumption credit are important for agricultural population

Poor people normally face difficulties in accessing credit, especially formal credit The reasons for loan application rejection may be due to the poor’s lack of collateral to ensure their loan repayment ability, imperfect information between them and lenders and high transaction and administrative cost Moreover, lending to the poor may be riskier since they may lose everything if they encounter unanticipated shocks such as health, natural disaster, etc Therefore, the big question

is whether the poor should be crowded out of rural credit market, especially formal or/and semi-formal credit? The answer should be no since it involves efficiency of capital problems

When the market is perfect, the allocation of credit will not rely on the wealth, but rather on the marginal return of capital In order to attract lenders, the poor without initial capital will be willing to accept higher interest rates However, when market is imperfect, it is difficult to eliminate such problems moral hazard and adverse selection (Armendáriz & Morduch, 2005) The formal lenders such as banks or financial institution can not ensure that the borrowers will use loans properly for right purpose, or repay the debt since the loan granted to the poor are

collateral-free Therefore, the formal lender will tend to charge very high interest rates However, the high interest rate will drive good borrower out of the rural credit market, especially formal sector

When being excluded from formal credit sector, the poor will seek for an alternative informal source of borrowing, for instance, relatives, friends, moneylenders, employers, and the rotating savings and credit associations (ROSCAs) In 1950s - 1980s, many governments in the developing countries attempted to intervene in the rural credit market by establishing rural state-owned banks to support farmers, especially the poor farmers (Morduch, 1999) However, heavy subsidies, political interests and corruption have inhibited the viability and success of these institutions Loans were also diverted away from the poor (Panjaitan-Drioadisuryo & Cloud, 1999; Armendáriz & Morduch, 2005) Therefore, microfinance is considered as a solution to credit constraint for the rural poor

2.2.2 Asymmetric information

The rural credit market plays significant role in agricultural production and rural development in developing countries Formal rural credit market is typically characterized by the disequilibrium of demand and supply, particularly excess demand Thenceforth, only a certain number of households who demand for credit can access to formal sources However, in reality, shortage of credit is not simply attributed to excess demand but to asymmetric information

Asymmetric information is a situation in which one party has more or better information than the other party The later may possess no information or wrong information about the goods or service in translation Theory of asymmetric information is initiated by Akerlof (1970) and then developed and researched by Spence (1978) and Stiglitz (1975) Asymmetric information may lead to adverse selection and moral hazard Asymmetric information theory is applied in many fields and researched by many authors This section presents this theory associated

with credit market Asymmetric information in credit market occurs when there is difference in information possessed by the lenders and borrowers

Adverse selection normally occurs before moral hazard; that is, when the

lenders accept to provide borrowers with the loan Borrowers often conceal the information that may cause disadvantages, such as real income, current debt, motivation of borrowing, etc This may lead to the fact that the households who are really in need of credit cannot access to the load or granted with an amount less than their demand Meanwhile, some households who are not really in need of credit can access to full amount The consequence may be bad to both parties: (1) The lenders may face risk of bad debt or their credit programs may not serve target customers, for instance the poor; (2) The borrowers cannot borrow from the credit programs

The adverse selection problem arises in the screening process where transaction costs are associated with distinguishing between good and bad borrowers This borrowing cost may be reflected and compensated by interest rate Greater interest rate may compensate for high transaction cost in small loans, which thereby drive out good borrowers Thenceforth, only potentially-bad borrowers can borrow This may drive out the target borrowers of the program, for instance, the poor who are normally unable to pay for high interest rate

Moral Hazard is associated with monitoring and enforcement mechanism

The moral hazard occurs when the borrowers may not take effort to repay their loan since they know the lenders share part of the liability (Pham & Lensink, 2007) Therefore, these lenders tend to link default risk to household related characteristics, particularly lender-borrower relationship and social norms As a result, only a proportion of rural households are unable to obtain credit or borrow an amount less than demanded from formal sources This credit rationing leads to the problem that many rural households face credit constraints regardless of their repayment capability (Armendáriz & Morduch, 2005) In other words, lenders decide whether they should grant the credit and how much to grant on the basis of information on

the borrowers that they have, thus not all borrowers will receive the loan that they apply for or receive the full amount applied

To address the problems of adverse selection and moral hazard arising from imperfect information between financial institution and the rural households, the lenders usually require collateral to loans The purpose of collateral requirements is

to improve creditworthiness and solve the incentive and enforcement problems (Klein et al., 1999) Land (especially residential land), house, durable assets or savings are always preferred collaterals when households apply to borrow from formal credit sources However, such collateral requirement becomes more stringent since the borrowers are normally resource-poor In some case, agricultural land or productive durable assets may indicate the rural households’ capacity of production, which thereby increase their creditworthiness and ability of repayment

Since one merit of microcredit programs is collateral-free, in South Asia, high interest rates and loan costs reflect information costs of dealing with adverse selection, monitoring costs to avoid moral hazard, and the default risk of borrowers (Adams et al., 2003) However, the higher interest rates are, the more probability of adverse selection and moral hazard effects it could lead to (Morduch, 1999; Cull et al., 2007) In order to resolve this dilemma, group-based lending practice has been implemented A merit of group-based lending is to increase the effectiveness of transactions, lower transaction costs, improve the outreach of microcredit programs

as well enhance loan repayment

Using asymmetric information theory, this study will answer the Research

question 1 In particular, this study focuses on whether the households of less

advantaged groups (e.g the poor, women, etc.) can have better accessibility to credit, as these groups are normally target clients of microcredit programs

2.2.3 Theories on accessibility to credit

Credit market follows the sequential two-stage lending process, where an agent (e.g household, firm) with demand for credit apply to borrow at the first

stage, then lenders will decide to choose whether and how much to grant the at the next stage (Aleem, 1990) Therefore, accessibility to credit, including microcredit, can be conceptualized as a sequential two-stage decision making process that is initiated at the demand side, and then followed by supply side (Zeller, 1994; Li et al., 2011a; Khoi et al., 2013) This research will discuss theory of choice and credit rationing theory and as “loadstars” for the analysis of accessibility to credit

2.2.3.1 Theory of Choice (Demand for Credit)

At the first stage (demand), an individual or a household wants to maximize his or her expected utility from the decision whether and how much to borrow (Zeller, 1994) Loan has its own opportunity cost which is reflected via the interest rate and; thenceforth, the individual or household’s decision on whether to borrow

or not is a rational choice based on demand theory

Discrete choice model is defined as a model that determines the probability

of a certain choice as a function of the utility derived from that choice (Cramer, 1991) Discrete choice analysis employs the principle of utility maximization (Ben-Akiva & Lerman, 1985) An agent will choose the option with the highest utility among those available choice set when making decision Accordingly, the probability that an option is selected is considered as the probability that the option may yield the greatest expected utility among other options (Train, 2009; Ben-Akiva and Lerman, 1985)

Supposed that there are two options a and b in a set of options (O m ) Let U ia and U ib denote the utilities that agent i gains from option a and b, respectively An agent will choose option i if U ia > U ib Thenceforth, follow (Li et al., 2011a; Ben-

Akiva and Lerman, 1985) the probability that agent i select option a from O m is

expressed using the following equation:

Pr (a| O m ) = Pr (U ia > U ib, ∀ a, b∈Om and a ≠b]

Similarly, if agent i choose option b, then the function is as follows:

Pr (b| O m ) = Pr (U ib > U ia, ∀ a, b∈Om and a ≠b]

It is noted that demand theory has been applied in many researches which investigate the choice of an agent Therefore, demand theory can be also useful for investigating the demand for credit (the first research objective) and households’ decision toward adoption of agricultural technology (second research objective)

For the first objective, households i will decide to borrow when utility from borrowing is greater than utility from non-borrowing Similarly, for the second objective, households i will decide to adopt when utility from adoption is greater than utility from non-adoption

2.2.3.2 Theory of Credit Rationing

Credit is a scarce resource due to the availability of credit Accessibility to credit is not simply explained by demand theory due to the fact that credit is rationed under asymmetric information (Stiglitz & Weiss, 1981)

Stiglitz and Weiss (1981) introduce the credit rationing theory on the basis two primary assumptions: i) lenders are unable to differentiate risk level associated with good (safe) and bad (risky) borrowers, and ii) loans are subject to the limited liability of borrowers to repay loans (i.e., if the returns from investment are less than the debt obligations, the borrower bears no responsibility in paying out of pocket)

Credit market does not simply the theory of supply and demand as typically described in markets of other goods and services Suppliers (i.e lenders) can not simply use market price (interest rate) to charge the clients (i.e borrowers) for the quantity demanded (amount of loan) In other words, lenders cannot simply increase equilibrium interest rate in the credit market because lenders do not have sufficient information on the risk that borrowers will default their loans Thenceforth, at the second stage, lenders tend to ration the loans that will be granted to borrowers In other words, lenders will make their decision on whether the loan should be provided, and then the amount to be granted, based on the lenders’ perception on the clients’ credit-worthiness or repayment capacity (Aleem, 1990)

Stiglitz and Weiss (1981) initiate credit rationing theory to explain the reason why some households can get the loan while the other can not Credit providers are able to judge the borrowers’ credit-worthiness based on the available information during the screening process Without sufficient information, information asymmetry may lead to at least two popular problems: adverse selection and moral hazard Adverse selection occurs during the screening process when lenders tend to favor “good” borrowers but crow out the target clients (Khoi et al., 2013) Moral hazard arises after the loan is granted The borrowers may not make effort to repay the loan since they perceive that the lenders may share the risk (Islam, 2007; Pham

& Lensink, 2007; cited in Khoi et al., 2013) In general, imperfect information will affect the lenders’ decision on whether and how much credit will be granted Thenceforth, not all households can get the loan or the amount that they demand

When analyzing credit rationing, it is important to know the households’ rationing status Feder (1990) define the borrowers who demand for more credit and the non-borrowers who could not obtain credit are classified as credit-constrained, and the remaining households are the unconstrained To examine rationing status of farmers, Jappelli (1990) employs a qualitative approach based on directly asking the applicants and non-applicants about their borrowing experience Applicants are categorized into such groups as: (1) If they desire to borrow more, which indicates

an excess demand, they are partially credit-rationed; (2) Applicants who did not

obtain any loans are classified as completely rejected As for the non-applicants, (3)

if they had intention to apply for credit but did not apply since their application

might have been rejected, they are classified as discouraged non-borrowers All these group are considered as credit-constrained Similarly, Baydas et al (1994)

and Zeller (1994) identify four group of respondents The constrained borrowers are

divided into such groups as: (1) completely rejected, i.e those who apply to borrow for a loan but unable to obtain; (2) unsatisfied, i.e those who apply and are granted with a loan less than demand The unconstrained include (3) satisfied, i.e those who

obtained full amount demanded; and the (4) non-applicants i.e those who do not

apply to borrow

Following Jappelli (1990), Feder (1990), Baydas et al (1994), Zeller (1994) Ali et al (2014), applicants and non-applicants can be categorized into five groups

Figure 2.1 describe in details the procedure to classify such groups

Some main reasons for credit constraints in rural areas are (1) Too high interest rate; (2) No collateral; (3) Fear of collateral loss; (4) No lenders; (5) Not know where to borrow; (6) Fear of being rejected; (7) Don’t like to be indebted or

go deeper into debt Since one merit of microcredit is collateral-free, the reasons for being constrained when applying to borrow from microcredit sources exclude (2)

No collateral; and (3) Fear of losing collateral

Figure 2.1 – Credit constraint and unconstraint

Using theory of accessibility to credit, this study will answer the Research

question 1 In this research objective, the sample is also divided into the poor and

the non-poor for further investigation The rational for this decomposition is that the poor is normally credit constrained than the non-poor, which may make the determinants of accessibility to credit have different effects among two group

Sample

Not Apply

to borrow

Apply to borrow

(Constrained)

Not desire to borrow

more

(Unconstrained)

2.2.4 Theoretical framework of outreach of microcredit

Assessment on the performance of microfinance organizations have been relied on the concepts of outreach and sustainability (Yaron, 1994, cited in Navajas

et al., 2000) Outreach and sustainability are expressed in terms of the theory of social welfare At first glance, outreach indicates the number of clients served However, according to Meyer (2002), outreach is multi-dimensional concept and

we need to look in to different aspects when measuring outreach Outreach is normally proxied by gender or poverty of borrowers, loan size, loan terms, transaction costs incurred by borrowers, number of borrowers, the financial and organizational strength of the lender, and the number of services provided by lenders Outreach is considered as social value of the output of microfinance institutions and measured by six aspects including depth, worth to users, cost to users, breadth, length, and scope (Navajas et al., 2000)

• Depth of outreach is "the value the society attaches to the net gain from

the use of the microcredit by a given borrower" This measure is to identify the poor or/ and women clients since society seems to value the net gain from a small loan for the poor or female clients more than the same gain for the rich or male clients The poor may have difficulty in accessing credit from formal sources since they have no collateral or fail

to signal their repayment capacity (Conning, 1999) Similarly, women normally face more difficulties than men in accessing the loan The number of the poor and women clients serves by MFIs may be considered

as criteria of outreach Therefore, poverty and gender are good proxies for outreach

• Worth of outreach to users refers to the amount that borrowers are willing

to pay for a loan Worth relies on the loan contract and tastes, constraints, and opportunities of borrowers More worth will mean more gain, given cost to borrowers constant