Impact of microcredit on the income of poor households in the southeast region

in some ASEAN countries is as follows: 10 In India, the standard is 2250 calories/person/day. In Bangladesh, the standard is 2100 calories/person/day. In Indonesia: In the early 1980s, the consumption of calories was 2100 calories/person/day as a standard to determine the boundaries between rich and poor. In China: In 1990, the consumption of calories was 2150 calories/person/day. Industrialized countries in Europe: 2570 calories/person/day. 2.2.3. Poverty line of Vietnam In the 1990s, the poverty line in Vietnam was defined as: households with income per capita in rural and mountainous areas are from 45,000 VND/person/month (540,000 VND/person/year) downwards. In rural and delta areas, the households with average income per capita was VND 70,000/person/month (840,000 VND/person/year). In urban areas, the income per capita was 100,000 VND/person/month (1,200,000 VND/person/year). In 2006, the poverty line in rural areas was 200,000 VND/person/month and in urban areas was 260,000 VND/person/month. In the period of 2011-2015, the poverty line was 500,000 VND/person/month (urban area) and 400,000 VND/person/month (rural area) 2.3. Income 2.3.1. Concept The General Statistics Office (GSO) (2011) defines that: Income is the total amount of money that a person or family earned in a day, a week or a month, or more specifically, all that a person earned when devoting the work force properly, so it is called income. The monthly income per capita is calculated by dividing the total income in year of the household by the number of household members and dividing by 12 months. 11 2.3.2. Factors affecting income The income of each individual is obtained from devoting the work force, participating in labor activities. The contribution of labor force of each individual brings income to each individual and family. That contribution brings economic value to the family through daily, monthly or even annually income for the household. Therefore, to create the valuable products, it requires a combination of physical capital and human capital (Ismail and Yussof, 2010). The combination of human capital and physical capital produces the labor value expressed by the added value of income achieved. In human labor activities, these two inseparable factors as follows. - Relationship between financial capital and income Capital is considered as a “lever” for the process of economic growth and development, a stimulation of the process of expanding the scale of production, implementation of economic projects and a contribution to increase the benefits, create the momentum to the economic development process. When credit markets are limited, the producing decision of a household depends on the price of market efficiency, p, the characteristics of production and the characteristics of access to credit shall be: q q( p q h K). - Difference in income The Neoclassical economists gave the theories: Human capital theory, income and discrimination theory, production theory ... to explain the fundamental differences in the income of individuals or households. Thus, the income is a multivariate function depending on many different factors (Y =f (x1,x2,x3xn). Today, to analyze the impact of factors on income, the function most commonly used in the analysis is the Cobb-Daughlas function. The Cobb-Daughlas function has the form as follows: Y = A.X1 α1 . X1 α2 . Xn αn .e β i D+x i D 1 +λ i D 2 12 In which, Y is the income; A is the constant; Xi ( 1, )i n is the independent variable affecting the dependent variable, Y, (income). The hypothesized independent variables include credit, household characteristics, environmental factors and related policies; e is factors other than factor Xi. In addition to the Cobb-Daughlas function, there is also a semi-logarithm function: LN(Y)=β0+β1X1+β2X2++βnXn+ei (Mincer, 1974). Or multivariate linear function such as Y = β0 + β1X1 + β2X2 ++ βnXn + ei is also widely used to estimate household income. From here, the study selects the proposes the research model that is a regression model with the form of multi- variable linear function to evaluate the impact of microcredit on income of poor households as follows: Y = β0 + β1X1 + β2X2 ++ βnXn + ei 2.3.3. Microcredit for income generating activities Microcredit is needed to help poor households generate income (Krog, 2000). Microcredit activities are used in developing countries and are highly effective in poverty reduction, especially microcredit focuses on women customers in rural areas, who have no accessibility capital resources from other financial institutions by barriers on collateral and complex and cumbersome procedures, helping them to create jobs and generate income (Mohanan, 2005). 2.4. Overview of access credit theory and barriers to access to credit 2.4.1. Asymmetric information in credit transaction and credit restriction Asymmetric information occurs when the borrower understands their ability to repay their loans while the lender does not know the borrower’s limitations. At the same time, the borrower does not collect enough information about the loan or the lending institutions. As a result, the borrower desires to borrow but cannot access to the loan; and the lender does not know the customers who need the loan and bringing about the 13 consequences associated such as lending under the relationship, other complicated and cumbersome procedures such as collaterals, guarantors, etc. But to meet these conditions, most small and micro enterprises, poor households and low income households cannot meet this condition. 2.4.2. Social capital, measurement of social capital and accessibility to credit Until 1990, the American sociologist, James Coleman, gave the concept that social capital as the characteristics in the everyday life, social networks, norms, and social trust helps members in the society to work together effectively to achieve the common goals. Bourdieu (1986) defines that social capital is derived from a direct or indirect network, and is a durable network of interrelated relationships that are be acquainted and recognize each other. 2.4.3. Characteristics of households, environmental factors and policies with access to credit Environmental factor with the distance on geographic gap, housing location and characteristics of the living area, and the inadequacy of imperfect information make it difficult for customers to access capital but they have to deal with difficulties to access (Le Khuong Ninh, 2016, Nguyen Trong Hoai, 2005). People who want to borrow cannot borrow and people who do not need the loans are more likely to get access, thereby creating distorting motives when borrowing, to get the loans, people who want to borrow seeks the way to get loans through the relationship, leading to asymmetry in the supply and demand of credit in the financial market. 2.5. Overview of relevant research documents 14 Table 2.2. Summary table of relevant research results Studies Object and scope of study Study method Study results Microcredit with income Quach Manh Hao (2005) Access to credit and poverty reduction in rural Vietnam With a cross-sectional dataset and econometric model analysis through the field survey data and the population living standard survey dataset VLSS 1992/1993 VLSS and VLSS 1997/1998 Characteristics of household affecting the access to credit and income of poor households Phan Dinh Khoi (2012) Living standards of poor households in the Mekong Delta With a cross-sectional dataset for survey in the Mekong Delta, the study used the Propensity Score Matching (PSM) technique and Instrumental Variable (IV-PE) method in the statistical analysis. The impact of microcredit on income has not been demonstrated clearly in this study. Phung Duc Tung et al., (2013) The work of poverty reduction in Ho Chi Minh City in period of 2009-2013 Quantitative and qualitative method with discontinuous regression model in the study Credit has impact on poverty reduction 15 Dinh Phi Ho & Dong Duc (2015) Income of poor households in Vietnam in the period of 2006- 2012 The study uses the Difference in Differences (DID) method to assess the impact of formal credit on incomes and expenditures of farmer households Credit has officially impacted on household incomes and expenditures. Mohanan (2005) Study the microcredit activities in India Qualitative research method Microcredit affects the income- generating capacity of the poor, empowering the women position. Islam and Ahmed (2010) Study the microcredit impact on income generating activities of customers Quantitative method with statistical analysis through SEM linear model Microcredit affects the ability to do business, create jobs, and build assets of customers Brown (2010) Study the microcredit impact on economic development, job creation and income generation Qualitative research method Microcredit offers borrowers the opportunity to increase the labor changes, diversify the livelihoods for poor households Ahmed et al., (2011) The role of microcredit in The qualitative analysis method with the analysis of random interview data Microcredit has no statistically 16 economic development and poverty reduction from 20 people surveyed significant impact Vitor et al., (2012) Microcredit with income of women in Central Ghana Quantitative research method with logit regression model, survey data from 300 microcredit women borrowers Microcredit helps improve the business skills and generate the income of credit women borrowers. Rykiye (2012) Study the microcredit with income change of participating members The survey data analysis method from 2,036 observations of poor households in Turkey MICROCREDIT does not have an impact on income, microcredit borrowers as a normal need for capital. Ayen (2016) Study the differences in income of microcredit borrowers and non- borrowers Quantitative analysis method with Propensity Score Matching (PSM) from the female head of household survey dataset at Jimma Zone There is a difference in income between borrowers and non- borrowers Accessibility to credit of households Nguyen Quoc Oanh and Pham Thi My Study the accessibility Quantitative method with two- step regression technique of The monthly income and purpose of borrowing affects accessibility to 17 Dung (2010) of formal credit Heckman (1999) through a survey dataset of 116 households formal credit of households. Nguyen Phuong Le and Nguyen Mau Dung (2011) Accessibility to formal credit resources of farmer households Statistical method describes the survey data from 60 households interviewed based on the pre-set questionnaire structure The families with better economic conditions can access higher credit Phan Dinh Khoi (2013) Formal and informal accessibility of farmers in the Mekong Delta Statistical analysis based on survey data of 358 observation samples from 919 rural households Income, employment of households affect accessibility of households. Tran Ai Kiet and Huynh Trung Thoi (2013) Accessibility to formal credit of farmers in An Giang Statistical analysis based on the survey results of 150 farmer households in the studied area The values of assets, income, and purpose of borrowing affect the accessibility to formal credit of farmer households. AFD (2008) Accessibility to microcredit of rural households in Morocco Causal analysis compares the ability to borrow from microcredit Rural households may participate more in microcredit loan when they have stable income and are not affected by seasonal factor. Ibrhim and Bauer (2013) Access to microcredit and the impact of the microcredit approach to Quantitative method with probit regression technique The characteristics of households with members with good production experience can access more 18 income of farmer households Masud and Islam (2014) Study the social capital with access to credit of households in Bangladesh Quantitative method with probit regression technique through random survey data from 153 households in Bangladesh Social capital affects the access to credit of households Source: Summaries of the author 19 2.6. Theoretical foundations forming theoretical frameworks for study and the establishment of research hypotheses Theory of capital resource of Ismail and Yussof (2010) shown that income is formed by physical capital and human capital. Physical capital is acquired by themselves or borrowing in the form of monetary or physical (physical is shown in the form of assets, labor materials, means of production, etc.). Human capital is gained through the labor accumulation, which is the skills and knowledge accumulated during the learning and life experience. During creating value for labor product, it cannot lack one of these two elements above; they complete each other and indispensable in every human activity. Human capital is involved in the operation, creation on the basis of physical capital and vice versa. Thus, when considering the income of an individual, a family needs a general assessment of the many factors that make up the value of income. Therefore, to assess the income of poor households, it is necessary to assess overall between the income and access of microcredit of households. Speeding up income for poor households on the basis of increasing access to microcredit for poor households. With the above arguments, the thesis establishes two research models and needs to be clarified in the study and theoretical framework of the study formed as Figure 2.2. Access to MC Income Household characteristics Household characteristics Environmental characteristics and non-financial activities Environment and preferential credit policies of the locality Social capital Microcredit 20 Figure 2.2. Theoretical framework for study Source: The author proposes based on theory foundations and previous studies Based on the above studies and based on the arguments of the research theory foundations, the thesis formed hypothesis on income of households affected by the following groups of hypothesis: (1) Microcredit (group of hypothesis H1); (2) Household characteristics (employment, number of labor generating jobs, number of dependents in the family - group of hypothesis H2); (3) Environmental characteristics and non-financial activities (group of hypothesis H3) Thus, it shown that, in order to promote the income increasing, it requires to increase accessibility to microcredit for them. The thesis establishes a Microcredit access model based on the foundations of the previous theory and studies. The summarized groups of hypothesis are included: (1) Social capital (group of hypothesis H4); (2) Household characteristics (income, employment - group of hypothesis H5); (3) Environment and preferential credit policies of the locality (group of hypothesis H6). 2.7. Gaps in research Through examine summarily the relevant studies, most studies affirm that microcredit offers many benefits to the poor such as increasing welfare, increasing empowerment for women, generating income and improving the living. By this argument, it shown that the gaps should continue to be inherited and clarified in the study are: (1) the study of the microcredit impact on the income of poor households and to find out whether there is the income difference between two groups of microcredit borrowers and non-borrowers? (2) the 21 specificity of the study area and (3) through the specificity of the area, to determine the impact of microcredit through the value of income and to increase the income and to increases the accessibility of microcredit for poor households, what limitations affects the accessibility of microcredit of poor households? From these gaps, the thesis continues to inherit previous studies and clarify gaps in this research thesis. 2.8. Conclusions of Chapter 2 Chapter 2 summed up the relevant theoretical foundations, examined summarily the studies in the country and abroad, and related studies; from that, the thesis has developed a theoretical framework for study. CHAPTER 3: STUDY METHODS 3.1. Study models As presented in the arguments of study theory in Chapter 2, the study uses a multivariate linear regression model with the hypothesis of microcredit impact on the change of poor household income and the Binary Logistic regression on the hypothesis of access factor to microcredit. 3.1.1. Microcredit model affects the income of poor households The thesis uses a multivariate linear regression model to test the study hypothesis. The analytical techniques through linear regression model, and the model has the following form: Y = β0 + β 1X1 + β 2X2+ ... + β 9X9 + e (3.1) e: Residuals β0: Vertical axis-cutting factor; β1 to β9: Regression coefficient (correlation) of the variable X1,..X9. 22 Dependent variable, Y: Per capita income of the poor household, measured by the total income of the household divided by the number of household members (unit: million/year). The model should go through the system with the following 6 tests: Firstly, Test the partial correlation of the regression coefficients Secondly, interpretation level of the model Thirdly, suitability level of the model Fourthly, phenomenon of collinearity Fifthly, test the autocorrelation of residuals Sixthly, Heteroskedasticity 3.1.2. Model of factors affecting access to microcredit Logistic regression model: ∑ (3.2) In which: Y: The dependent variable has two states (0,1); X1, X2Xi is the value of independent variables; β0 is the estimated value of Y when the variables X value 0; βk is the regression coefficients; u is residual. According to Cox, D.R (1970), the general form of the Binary Logistic regression model is as follows: = β0 + β1X1 + β2X2 + βnXn (3.3.) In which, P (Y=1)=P0: Probability of households access to microcredit; and P(Y=0)=1-P: Probability of household not access to microcredit. [ ] 23 Take 00 = ; with 00: Odds Coefficient LnO0 = β0 + β1X1 + β2X2 + β7X7 (3.5) Therefore, Log of Odds coefficient is a linear function with independent variables Xi ( 1,7)i  . Equation (3.5) has the form of a logit function. According to Agresti (2007), the model is approved the test system includes: Firstly, the Wald test. Secondly, test the suitability of the model. Thirdly, test the level of interpretation of the model. 3.1.3. Developing the basis for variable selection in study models 3.1.3.1. Microcredit model affecting income (M1) Figure 3.1: Model of impact of microcredit on household income Source: Propose to study based on theoretical framework and previous studies MICROCREDIT (hypothesis H1) HOUSEHOLD CHARACTERISTICS (hypothesis H2) Microcredit is expressed through factors: loan scale, interest rate, loan term and loan purpose Employment Environmental risks Number of dependents Non-financial policies ENVIRONMENTAL CHARACTERISTICS AND NON-FINANCIAL SUPPORT POLICIES (hypothesis H3) INCOME Labor scale 24 Selected variable Hypothesis Basis for variable selection [X1]. QM_VON H1 Bateman’s Theory (2010) and Janvry’s Theory (1995) Brown’s studies (2010); Islam and Ahmed (2010); Vitor et al., (2012); Ibrahim and Bauer (2013), Banerjee and Dulfo (2016); Alhassan and Akuduga (2012). Vitor et al., (2012). [X2]. TH_VAY [X3]. L_SUAT [X4]. MD_VAY [X5]. S_PTHUOC H2 Dinh Phi Ho and Dong Duc (2015) [X6]. QM_LDONG World Bank (2012); Ismail and Yussof’s Theory (2010); Dinh Phi Ho and Dong Duc (2015) [X7]. V_LAM World Bank (2012); Ismail and Yussof’s Theory (2010); Dinh Phi Ho and Dong Duc (2015) [X8]. R_RO H3 Janvry’s Theory (1995), Dinh Phi Ho and Dong Duc (2015) [X9]. CS_TPC Manganhele (2010), Phung Duc Tung et al., (2013); Banerjee and Dulfo (2016); Boamah and Alam (2016); Nguyen Duc Nhat et al., (2013); Alhassan and Akuduga (2012). Source: Summarize from theory foundations and previous studies 3.1.3.2. Microcredit access model (MH2) 25 Figure 3.2. Microcredit access model Table 3.2. Summarize the foundations for developing variables for the microcredit access model Selected variable Hypothesis Basis for variable selection [X1]. VON_XH H4 Masud and Islam (2014); Putnam (1995); Baurm and Ziersch (2003); Stone (2001); Kilpatrick (2002); Ajam (2009); Lin et al., (20010); Okten (2004). [X2]. TS_TGVXH [X3]. V_LAM H5 AFD (2008); Tran Ai Ket and Huynh Trung Thoi (2013); Phan Dinh Khoi (2013). [X4]. T_NHAP Brown (2010); Armed et al., (2011); Vitor et al., (2012); Ibrahim and Bauer (2013) Mohannan (2005), Phan Dinh Khoi (2013). Tran Ai Ket and Huynh Trung Thoi (2013). Household characteristics (hypothesis H5): Income, employment Social capital (hypothesis H4): Participating in social capital, frequency of participating in social capital Environmental factors and preferential credit policies of the locality (hypothesis H6): The housing location of the households, living area and preferential policies of the locality Access to microcre dit 26 [X5].V_TRI H6 Le Khuong Ninh (2016); Dinh Phi Ho (2012), Banerjee and Dulfo (2016) and characteristics of the study area. [X6]. K_VUC World Bank (2012), Le Khuong Ninh (2016), Nguyen Trong Hoai (2005). Phan Dinh Khoi (2013) and characteristics of the study area [X7]. CS_TC Phung Duc Tung et al., (2012), Boamah and Alam (2016); Nguyen Duc Nhat et al., (2013); Alhassan and Akuduga (2012). Source: Summarize proposals from study theory foundation and previous studies 3.2. Measurement of concepts in study models  Model 1: The impact model of microcredit on income change Table 3.3: Measurement of variables in the model of microcredit impacts income No. Contents Measurement Sign expectat ion I INDEPENDENT VARIABLES: 1 [X1]. QM_VON Representing the amount of loan (Unit: mil dong) + 2 [X2]. TH_VAY Term of loan use (unit: month) + 3 [X3]. MD_VAY Loan purpose + 4 [X4]. L_SUAT Interest rate by month (unit:%/month) - 5 [X5]. R_RO The risks in the past 3 years of the household, value 1 = yes and vice versa = 0. - 6 [X6]. S_PTHUOC Number of children and adults outside the working age (children under 15 years old, older than 60 years for men and 55 years for women), (unit: person) - 7 [X7]. QM_LDONG Number of main employees in the household generating income for the family during the past 6 + 27 months (unit: person) 8 [X8]. V_LAM Employment of the main decision maker/ householder in the family, value 1 = employment (within 6 months) and vice versa = 0. + 9 [X9]. CS_PTC The support policies for non-financial activities such as creating a livelihood environment to help the poor create jobs, train to transfer of science and technology knowledge, cultivating and breeding knowledge, develop the business plans, labor and occupation, ... value 1 = have support policy and vice versa = 0. + II DEPENDENT VARIABLE (Y): The income per capita is measured by the total income of the poor household divided by the number of household members (mil dong)  Model 2: Microcredit access model Table 3.4. Measurement of variables in the study of microcredit approach study model Variable name Sign expectation Measurement Dependent variable: Microcredit Binary variable, value=1 if it participates in microcredit borrowing and value=0, by contrast. Independent variables: [X1].VON_XH + Households participating in activities in the locality, value = 1 and vice versa = 0. [X2]. TS_TGVXH + Number of times (frequency) participating in activities/last 6 months (unit: times) [X3].V_LAM + Householder or main decision maker in the household, value = 1 means to have employment and vice versa = 0 28 [X4].T_NHAP + Income per capita/year (mil dong) [X5].V_TRI - Housing location of the household compared to the main road (inter-commune traffic) (unit: km) [X6].K_VUC + Living area of household: Urban area, value = 1 and vice versa = 0. [X7].CS_TC Preferential financial policies of the locality, value = 1 means get preferential poli