INTRODUCTION

1. Background and relevance of the topic

During the 1990’s, the provision of micro-finance has increasingly been acknowledged as an effective means for poverty reduction in many developing countries. For the poor, lack of financial security is one of the critical reasons for their continuing poverty. Despite the merits of a policy providing the rural poor with subsidized credit, the provision of such credit often meets only a small fraction of the total credit needs of poor households. What is not met through official channels is normally met through the informal financial market.

Vietnam is one of the poorest countries in the world. The poverty incidence in Vietnam is still high. Based on the general poverty line determined by the World Bank (WB) and General Statistics Office (GSO), the poor were accounted for 29.7 percent of population (Vietnam Living Standard Survey 1997-1998, VLSS98). Causes of poverty in Vietnam have extensively been examined. It is found that "limited access to available resource, including financial capital, is among the main underlining causes of poverty in Vietnam. Microfinance is a key part of such resources from which the poor can choose and develop better livelihood" (UNDP, 1996).

To date, there have been several studies of general problems faced by the rural poor households in assessing formal credit in Vietnam that the rural households' ability to access formal credit is limited by complicated procedures of formal financial organizations while the informal sector with flexible conditions makes it easier for farmers to borrow, including Tran. T. D. (1998), Tran. V. B. et all (1999), Le. T. T. et all (1999), Do. Q. T. (2000). However, less work has been done on analyzing the factors affecting the rural poor’s accessibility to credit using data from a nation-wide household survey. Therefore, an investigation into the determinants of rural households’ access to credit in the formal and informal credit markets is worth undertaking.

2. Focus and scope of the thesis

The aim of this thesis is to examine the factors affecting the rural poor’s accessibility to credit market. In particular, this research analyzes qualitatively and quantitatively the factors affecting accessibility of the rural poor to credit in Vietnam in 1997/1998. The study concentrates on the rural poor. Data from the rural non-poor is also used for comparing and distinguishing the salient characteristics of rural poor borrowers.

Ø      Primary data: obtained from VLSS98. This survey was implemented by GSO with funding from Swedish International Development Cooperation Agency (SIDA) and UNDP, and technical assistance from WB.

Ø      Secondary data: collected from various sources, including VLSS93, microfinance survey conducted by Master in Development Economics (MDE) team, Department for International Development (DFID) and State Bank of Vietnam (SBV), policy statements, official unofficial reports, various comments and figures from published studies in the field, newspapers, reports of conferences and documents and papers on the Internet

3. Research questions

Research question: "What are the factors affecting the rural poor’s access to (formal/informal) credit in Vietnam?", and sub-questions:

?   What are the characteristics of formal and informal borrowing by the rural poor?

   What is the difference between the rural poor’s access to formal and informal credit?

?   What can be done to improve the rural poor’s access to formal credit in Vietnam?

4. Structure of the thesis

The paper consists of an introduction, four chapters, a bibliography and appendices. Chapter 1 represents a literature review of three main approaches to credit market, then basic model explaining the behavior of borrowers. Chapter 2 provides a detailed descriptive analysis of borrowing by rural poor with an emphasis on characteristics of the poor’s loans with important findings through the analysis of rural credit data. In chapter 3, the simultaneous equation models are used to test hypotheses in the beginning of the study as well as suggested in examination of household credit data. Chapter 4 provides a conclusion and suggests some policy implications for further development of rural finance, and improving rural poor’s accessibility to formal credit.

CHAPTER 1 – FORMAL VERSUS INFORMAL MICROFINANCE -                   A LITERATURE REVIEW

This chapter reviews three theoretical paradigms in contemporary financial economics providing analytical frameworks for explaining how the rural credit market works. This objective is to present a framework that incorporates the realities of the environment in which formal credit institutions and informal credit sector are operating and the rural poor are making decisions.

1. Three theoretical approaches

1.1. Traditional approach

The traditional approach views limited capital fund as the main factor impeding the access of farmer to formal credit institutions in developing countries. It was extension of Keynesian views on the government role formed during 1930s, and other economists such as Wai (1956), Cairncross (1962), Bottomley (1964). The traditional school based on its explanation for the supply of and demand for credit. From the supply side, they assumed that low levels of income limit the savings potential in developing countries. The role of the government in increasing savings, creating credit, and providing incentives for the proper sectoral allocation of the limited loanable funds, as a result becomes crucial.

From the demand side, credit is considered as input. Traditional approach assumed that the borrowers are more sensitive to interest rates than lenders. Low interest rate helps farmers to make productive investment and use new technology (Wai 1956, 1957; Bottomley 1964). In this regard, the traditional school advocated cheap credit policies implemented through interest rate ceiling, anti-using laws and interest rate subsidies. The traditional approach implies a high level of government intervention in the form of targeted credit.

Ø      Weakness of the traditional approach: Subsidized credit programs, in sum, tend to succeed all too well in keeping governments in power through political patronage and in maintaining and even enhancing, the position of rural elite. Contrary to their intent, credit policies have often been biased against the poor. Low return of subsistence food production as insurance in a risky environment is the situation of the poor. That is argument of the Financial Repression Approach, which is discussed in next section.

1.2. Financial Repression Approach

While the traditional and financial repression school perceived credit markets as fragmented and imperfect, the financial repression approach argued that this is more a result of government policies which have repressed the growth of credit markets rather than an inherent characteristic of the market itself. Cheap credit in fact does not favor the poor, but instead allows the rich people to get a disproportionate share of the cheap credit. Faced with these disappointing results, the early 1970s, a large number of evaluation studies were undertaken, which challenged the traditional view, such as Donald (1976), Adams (1971, 1984), Gonzales-Vega (1976), Vogel (1981), Pischke (1981) and Ladman (1981, 1984).

The financial repression hypothesis raises issues of poor policies. Many policy-makers, technicians, writers on development often don't think interest rates as incentives or prices, and they fail to recognize the importance of these prices in affecting the behavior of participants in financial market. Next, the financial repression approach indicates the weakness of formal financial institution. In rural areas, operations of formal financial institution were largely restricted to state owned development banks and officialized cooperatives. Both were "used and abused" to channel priority credit with controlled and subsidized interest rate, to target credit according to political rather than banking criteria, and enjoyed generous credit guarantees to cover anticipated losses.

In addition, the finance repression approach sees the discrimination against the poor in formal finance sector. According to the Iron Law of Interest Rate Restrictions, as the loan portfolios of formal financial institutions usually include both rationed and non-rationed classes of borrowers, interest rate ceilings become more restrictive. The size of the loans granted to the non-rationed borrower classes increases, in contrast to diminish the size of the loans granted to the rationed borrower classes (Vega, p.86).

The fact was that formal lenders were either unable or unwilling to solve the information problems involved in the broad range of rural credit transactions and those high interest rates reflect high information costs, not the scarcity of funds. The critique from Pischke, Adams, and Donald (1983) was enlivened by observations of credit policies in developing countries. They stressed the distortion introduced by government policies and tended to idealize the informal credit markets that did exist or that might have existed in the absence of the massive government intervention in the credit market. There was a presumption that an intervention-free rural financial market would approximate the perfect competition model.

The financial repression approach suggests that financial liberalization would lead to financial deepening; improved efficiency, resulting in lower spreads between borrowing and lending rates, and increased flow of funds between segments, including better access to formal finance for previously marginalized savers and borrowers.

Ø      However, many surveys of formal and informal institutions and borrowers such as in Ghana, Malawi, and Tanzana investigate the hypothesis that reforming financially repressive policies would not be sufficient to overcome fragmentation of financial markets because of structural and institutional barriers to interactions across different market segments. The financial repression approach does not address the issues of incomplete markets and imperfect information in the context of credit market in rural areas. In much recent theoretical literature, the problems of moral hazard and adverse selection are assumed to be decisive for the organization of agrarian institutions. That’s why financial repression approach could not apply their literature to overcome difficulties in rural credit market.

1.3. The New Institutional Economic Approach

The institutional approach based on cost and incentive problems that emerged especially in a contract involving the promise to pay in the future. It argued that other approaches to credit market ignored the transaction costs beyond the interest rate. In the view of the institutional approach, these transaction costs limit the access of the poor to the formal credit market even though interest rate is low. A strand of literature beginning with Spence (1973), Rothschild and Stiglitz (1976) and Wilson (1977) stresses the role of "sorting devices" that facilitate the transmission of information from informed to uninformed agents in the markets. Then, Hoff and Stiglitz (1990) advanced an explanation based on imperfect information causing the issues of credit market.

Firstly, the terms of the loan contract such as collateral requirements may influence the characteristics of those who present themselves for loans and, hence, the distribution of the lenders' returns. If the lender cannot at reasonable cost distinguish good borrowers from bad, he face a potential problem of "hidden information'', to use Arrow's (1985) terminology - or as it is usually called adverse selection. Secondly, characteristics of the borrowers are fully unknown to the lender, the terms of the loan may influence the borrower's activities and performance in ways that affect the lender's returns. This causes moral hazard. Thirdly, transaction costs that the costs associated with the steps that borrowers undertake to complete the requirements of borrowing also lead to credit rationing. For any lender the larger size of transaction costs, assuming a fixed interest rate, the greater will be rationing power of transaction costs.

Transaction costs have an important impact on the structure of financial markets in rural areas where low average returns and high risks associated with many agricultural activities. There are also comparative advantages of finance formal and informal by strengthening the links between them. Remote areas with poor transportation may lack access to formal financial institutions. Meanwhile, informal sector relies on localized, personal information that gives them local monopoly power but constrains their ability to scale up (Stiglitz, 1992). An additional barrier to access is the cumbersome administrative procedures that accompany formal lending, raising borrowers' transaction costs to levels that are not much different from total costs of borrowing in the informal market.

Ø      If structural and institutional constraints are important, reforms in the formal financial sector would have little impact on informal activities, which would respond more to changes in financial demand and supply in the real economy than to change in financial policies. Therefore, we investigate further the demand for credit and rural poor’s borrowing behavior in the next section.

2. The behavior of borrowers and the demand for credit

2.1. Credit rationing and ceiling interest rate

Credit rationing is defined as a "situation in which the demand for commercial loans exceeds the supply of these loans at the commercial loan rate quoted by banks." (Cosci, p.7). Interest rate acts as screening which regulates the risk composition of loan portfolios, and as dual function of price and instrument for regulating risks. When excess demand happened, the price increased and set off excess demand. Thus, interest rates increased, making greater risk then offset banks' income increase from higher interest rate. The banks keep low interest rate and ration available funds, resulting in credit rationing with no-tendency for interest rates to rise.

A lower interest rate than the market equilibrium fixed at ie leads to shift the supply curve. At the ceiling interest rate ir while lenders could provide loans at Q's, potential borrowers would be willing to pay a higher interest rate i' for this limited amount of funds. Borrowers will therefore continue to seek credit equivalent to Q's as long as their transaction costs are less than or equal to margin i'ir. The lower the restricted interest rate, the greater the transaction costs that borrowers will be willing to absorb and vice versa.

2.2. Borrower behavior and the Demand for Credit

Like the first-stage of the two-stage model of Zeller (1994) discussed in details in next part of the thesis, the model that captures the reality that farmer-borrowers concern themselves with the total borrowing costs is suggested by Ladman (1984). Ladman began the model by studying the borrowing costs BC that are imposed by the lender's credit delivery system. Costs include interest costs IC and constant borrower transaction costs BTC . The condition of profit maximization is i = MRR. Therefore, the profit equation of the farmer-borrower is given by Ladman as in this form:

where:              AR = R/L                     is average revenue

                        ABC = i + BTC/L        is the average total borrowing costs.

As Figure 1.2, the farmer would want to borrow L*, where i = MRR and p would be L*.(AR - ABC). The figure shows that T1 is the borrowing threshold below which the borrower would not borrow from a lender.

Ladman's model is extended to two lenders in the market: one is a formal credit institution with a low nominal interest rate but high transaction costs for the farmer; one is an informal moneylender who charges a high interest rate but low transaction costs. The farmer therefore will choose the lender that offers the larger expected profit p given interest rates ii and if, transaction costs BTCi and BTCf, and demand for credit D, subject to the constraints and associated risks of the out-of-pocket expenses T2i and T2f (subscripts i and f are used to denote variables related to the informal and formal lender, correspondingly).

Ø      From this model for case of two lenders, Ladman suggested that offering cheap credit does not necessary induce farmer to borrow from formal lenders. Otherwise, the farmer's decision to borrow from whom is dependent on the total cost of borrowing. The borrowers seeking small loans will often prefer to work with lenders who charge high interest rates but who impose low transaction costs upon borrowers. When seeking larger loans, borrowers may prefer to work with lenders who impose larger transaction cost but charge a lower interest rate. Therefore, the model shows that rural credit market is characteristics of segmentation.

CHAPTER 2 – AN ANALYSIS OF VIETNAMESE RURAL CREDIT MARKET

1. General picture of the rural poor borrowers

This section presents the overview of the rural poor borrowers in terms of the percentage of borrow by the number of households in comparison with rural poor non-borrowers, non-poor borrowers.

Source: Vietnam Living Standard Survey, 1997-1998

The extension of formal banking system to rural Vietnam is most remarkable achievement of microfinance in Vietnam since 1996. By 1997-98, the percentage of rural households in debt had risen to 50 percent (54 percent of rural poor). Figure 2.1 shows that the poor are almost as likely as the non-poor households to borrow. Similarly, the average loan size of the poor is about 1,684 thousands dongs, which is much small in contrast to the average loan size of rural non-poor households 3,743 thousands dongs (see Table 2.2).

2. Formal and Informal funding sources existing in parallel

Even though there have been many official initiatives in the rural banking sector during the past few years, the vast majority of rural credit needs are still not being met by the formal banking sector. This is proved by the high rate of borrowing from informal sector. Table 2.1 shows the source of borrowing by the rural households. The table reveals two interesting points. First, a big gap is obviously shown in the average loan size of the rural poor and the rural households in both formal and informal sector. Government banks appear to be the first source of loans in terms of amount, however, it represents only 25.6 percent of loan sources when we account for number of loans.

Table 2.1 - Loan and the percentage of number of borrowers by sources

Source: Vietnam Living Standard Survey, 1997-1998

Second, for the rural poor, the informal sector supplies smaller loans than the formal sector did. Therefore, the proportion of informal loan values is lower than that of formal loans (accounted for 37.8 percent of the rural poor’s loan values) even though the rate of numbers of informal loans is more than half of the total numbers of loans. Yet, 54.3 percent of all poor borrowers had at least one active loan from the informal sector. We can therefore observe that the informal credit sector is still exist parallel to formal financial sector.

3. Characteristics of loan uses by rural poor: Formal versus Informal loans

As pointed out in Chapter 1, the formal sector institutions provide loans with priority for production in most credit contracts. The development of formal sector services has caused greater segmentation. In Vietnam, the formal lenders, notably the VBA and the VBP, ration under-priced credit to the richest and the poorest farmers and for investment purpose only. The lengthy approval procedures for formal loans also mean that immediate funding needs are met informally. The poor more likely to be victims of usurious lending, as they seek immediate consumption assistance in times of disasters, food or health crisis. Government banks provide loans for investment purpose only and thus lending for consumption purpose is left to the informal sector.

Table 2.2 – Loan uses by the rural borrowers

Source: Vietnam Living Standard Survey, 1997-1998

Table 2.2 contrasts the purposes of loan obtained by the rural poor with the loans obtained by the rural non-poor households. The data shows that the poor can borrow only small loans, especially for investment and working capital in comparison with the non-poor’s loans. It means that the poor have no other activities to diversify their investment to avoid risks. This might be one of reasons for high default of the rural poor’s loans. The proportion of consumption loans, including loans for residential construction, is much higher in informal loans, particularly from relatives and friends. Such variation reflects the differential ability of lenders to access the repayment capacity of borrowers. However, the data from VLSS98 show evidence of a diversion of loans to unapproved uses or called fungibility of credit. But it is not easy for the poor to do like this since they often lack social influence or relationships. That is why many the poor accept usurious interest rates. In particular, the poor may require loans from private moneylenders for consumption in the months immediately before harvests.

4. Characteristics of the rural poor’s loans: Formal vs. Informal lenders’ requirements

a) Collateral requirement:

One of "rationing mechanisms" that most formal institutions use is collateral requirement. To secure loans, formal lenders usually require physical collateral in the form of land, a house, or durable goods. Figure 2.2 brings a brief description of loans borrowed by rural poor under the condition of collateral and non-collateral.

Most formal loans require collateral, accounted for 28 percent of total number of loans borrowed by rural poor (equivalent to 73.2 percent of total number of loans with collateral). In contrast, loans without collateral are mainly provided by the informal sector, represented around 44 percent of total number of loans borrowed by rural poor (equivalent to 71.4 percent of total number of loans without collateral). According to VBA’s report, the bank provides a loan based on estimating capacity to payback loans, that loans is less than 70% of assessed value household assets, which serve as collateral.

Ø      Form of collateral: The most common type of collateral is land. Only about 30 percent of households have "red certificates" of land use rights. For more detail study of loans with collateral, the table gives the picture of how the different sources require different forms of collateral. As Table 2.3 represented, of 512 numbers of loans, 262 loans used house as collateral, 203 loans used land, however, the percentage of loan value in case of mortgaging land is more than half of total loan value (55.6 percent). The government banks supplied the largest proportion of loans with collateral, of those loans they required mostly land as collateral (153 loans over 272 loans). Meanwhile, other sources such as relatives, ROSCA, Poverty programs mainly used house as collateral for their loans.

       Source: Vietnam Living Standard Survey, 1997-1998

b) Interest rate

Figure 2.3 brings a general picture of providing loans to rural poor with and without interest. Interestingly, most loans provided by formal loan required paying interest while a large part of informal loans did not required interest. In microfinance, as elsewhere, the poor always suffer and have less favor. With less capital, less education and less income, they are a greater lending risk and hence, face higher interest rates especially from informal lender. Table 2.4 shows more detail the big different between two sectors: the informal lenders charged high interest rate in contrast to the low interest rate of the formal financial institutions. Easily, the difference can be explained by State Bank of Vietnam's policy on controlling ceiling interest rates and by government subsided credit policy for targeted groups. Interest rates of informal financial sector freely adjusted in credit market while interest rates of formal financial sector are sticky under market clearing rate.

  Source: Vietnam Living Standard Survey, 1997-1998

Table 2.4 - Monthly Interest rates by source of loans (percent)

  Source: Vietnam Living Standard Survey, 1997-1998

c) The size of loans

The other factors resulting in segmented market are regulations about the length and size of loans. Setting a maximum loan size that is very small is a mechanism for "self-selecting" poor households (as richer households may balk at the high transactions cost for a small loan). From Table 2.5, we can see the fact that the rural poor borrowed largely below 3,000 thousands dongs. In which, almost of loans provided by the informal lenders are very small loans (48.1 percent of loans below 500 thousand dongs) while the numbers of formal loans is popular from 2,000 to 3,000 thousand dongs. For large loan value (from three million dong up to ten million thousand dong), the government banks are the major lenders (66.4 percent), following are Relatives and Friends (14.0 percent). This result reinforces the conclusion that most rural poor borrowed with small loans, especially for petty loans from informal sector.

Table 2.5 - The number of loans divided by loan size and source

   Informal                                      349                164               175                  35                   2                 725

1. Moneylenders                                  66                 31                       37                   8                   1                 143

2. Relatives                                    141                  85            85                   21                                           332

3. ROSCA and others                      142                     48                     53                     6                       1                  250

   Formal                                          73                      167               249                   117                      4                  610

4. Poverty Programs                            26                   81                94                   12                               213

5. Government Banks                          35                   72               131                     101                   4                       343

6. Private Banks                                      7                   6                 10                           2                                            25

7. Other Programs                                  5                   8                 14                    2                                         29

Source: Vietnam Living Standard Survey, 1997-1998

5. Dispersion of borrowing by rural poor across the country

Regional differences can be seen clearly in Table 2.6. In some regions, particularly in the Red River Delta and Southeast, the government banks are relatively unimportant for the rural poor. Government bank loans are most common in the Mekong River Delta and Central Highlands. The numbers of loans that the rural poor can borrow from the formal lenders are less than that of the rural non-poor. Except in Northern Central Coast, in other regions across the countries the non-poor can take loans from formal sector more than the poor can. Of 1335 rural poor’s loans, 45.7 percent is to borrow from formal institutions that is lower than formal loan rate borrowed by rural non-poor.

             Source: Vietnam Living Standard Survey, 1997-1998

Apart from, there is a big gap of access to credit between the poor and non-poor in the same region, especially in Red River Delta and Southeast. The poor in Southeast only had 35.2 percent of loans to borrow from the government banks, private banks or state programs while the non-poor in this region can access to formal financial institutions up to 56.2 percent. Rural credit markets have always been segmented to some degree because of geographical constraints and weak information system.

6. Poor women and credit

This last section of the chapter is to mention about the issue of providing credit for women. Although much rural microfinance in Vietnam is channeled through women, the vast majority of rural loans are made to men. VLSS98 data indicate that women do not play the same role depending on the source of credit used. In most cases, it is men, as heads of the family, who decide and negotiate loans form moneylenders. Only 16.3 percent of loans were to women.

Ø      In brief, the rural credit market in Vietnam is a repressed financial sector. Demand for cheap loans outstrips supply. Rationing mechanism to allocate the credit are biased against the poor. The poor tend to be pushed into market segments where they have no formal sector lending options. Those with access to cheap credit may find themselves in the comfortable position of on-lending at higher interest rates to those without such access. To put our investigation further, the next chapter will specify statistical models to identify determinants of borrowing by the rural poor.

CHAPTER 3 – MODEL SPECIFICATIONS, EMPIRICAL RESULTS AND INTERPRETATIONS

1. Econometric framework & variables in simultaneous – equation models

According to Zeller (1994), a two - stage decision process represents how demand and supply side factors affect borrowing by households: at stage one, the household or its member decides whether to apply for credit or not; at stage two, it is screening process, the lender decides whether to give the applicant all the credit he or she asked for, or partially reduce the credit amount, or to fully reject his or her demand. We investigate in the first stage of Zeller’s model as getting loan rather than as applying a loan since there is limit of data of VLSS98. Probability model is used to estimate the determinants of probability of borrowing: Borrow (0 = not borrowed, 1 = borrowed). The second - stage model is used to estimate the determinants of the loan size. If we run separately each model, the results may be bias due to the second model lack of information from observation of the first model (such as information from those who not borrow) and vise versa. It is better to lump together a set of variables that can be determined simultaneously by the remaining set of variables – precisely what is done in simultaneous – equation models, as follows:

Probability to borrow:             Y₁ = f (HH, HE, R, U)                         (3.1)

Loan size:                               Y₂ = f (HH, HE, L, R, Y_1,U)                (3.2)

The dependent variable Y₁ is a dummy variable representing whether a rural poor household borrows or not while the dependent variable Y₂ is amount of loan borrowed by rural poor. The independent variables include sets of variables:

a) The vector of household head's characteristics HH includes age (AGE), gender (SEX), education (EDU) and social responsibility (SOCIAL). Under of Confucian philosophy, the head of household plays an important role in decision-making related to family as well as community. The characteristics of household head, therefore, are employed into model, to examine the hypothesis that the effect of household head’s age, sex, education and social responsibility.

b) The vector of household' characteristics endowments affecting credit demand - HE- includes 4 types of variables, including household size or the number of household members, namely HHSIZE, the number of wage earners and number of employed members, called Occ_wage and Occ_self respectively, total expenditure of rural poor, namely EXP.

c) The vector of loan characteristics - L: As mentioned above, due to high transaction costs, formal sector loans are normally large, while small borrowers like rural poor rationally seek loans from informal sources. That is reason for introducing the vector of loan characteristics, including the loan use (HB, PRO), loan source (formal or informal sector) and loan collateral requirement (COL).

In addition, the dependent variable in the first-stage model - Y₁ is put into the second – stage model to measure the effect of probability of getting credit to loan size. It is hypothesized that the higher probability of taking a loan, the higher amount of loan borrowed by rural poor.

The final set of variables R concerns regional location. Using this set of dummy variables of regions we are expected to find out which region the rural poor would have higher probability and borrow larger amount of loan.

Briefly, the simultaneous equation models with expected signs of variables are summarized as follows:

Table 3.1 – Expected signs of variables of the simultaneous equation models

2. Interpretation of Results

Coefficients of the simultaneous-equation models are estimated with data of 1,658 rural poor households estimated from VLSS98. The two-stage least squares (2SLS) method is usual alternative. The results of these models are represented in Table 3.2. The hypothesis that all coefficients equal zero is rejected. Although the coefficient of determination (R² =0.355 and 0.407 respectively to Y1 and Y2) in the simultaneous-equation regression is low but this is not unusual for studies using large number of variables as well as cross sectional data.

Table 3.2 – Simultaneous equation models:

Determinant of Probability of borrowing and loan size borrowed by rural poor:               

*:  significant at the 1 percent;           **: significant at the 5 percent;         ***: significant at the 10 percent

Ø      In the characteristics of household head - all the coefficients have the expected signs except the negative sign of variable AGE. The unexpected result shows that the young engages in borrowing more than the older, or an increase in the age of rural poor head of 1 percent, followed by a decrease in the probability of taking a loan about 14 percent with significant level at 1 percent. However, positive sign of Log(AGE) variable in second-stage model indicates that the older borrows larger loan size than the young.

Ø      The same positive sign of coefficients of sex in both equations shows that male heads of households participate in borrowing more as well as taking larger loan than female heads of households do. If household head is a man, his probability to get a formal loan may increase 6 percent and his loan size may increase by 215 thousand dongs at the 5 percent significant level.

Ø      As expected, schooling years of household head – EDU – has a positive influence on probability to borrow and loan size. The regression results indicate that the better-educated household head the more chance to take loan and larger loan size to be borrowed.

Ø      SOCIAL variable - having good relationships or position on community helps households borrow more. If rural poor have social influence, loan amount borrowed rises by around 167 thousands dongs at the 95 confidence level as well as increases by 3.5 percent of probability to get loans at the 10 percent significant level.

·        For the household's characteristic endowment, the HHSIZE variable  has unexpected signs on both models. The result implies the more members household has, the less chances as well as loan size household could borrow. In figure, if the household adds one member, it may be likely to decrease by 1 percent of probability of borrowing and reduce by 48 thousand dongs of loan size at 90 percent confidence level.

·        The signs of Occ_wage and Occ_self variable are negative, especially coefficient of Occ_wage variable has insignificant explanatory power. In comparison to occupation of household members in agriculture, the more self‑employed members the less likelihood of rural poor borrows.

·        Proxy for outflow of expenditure of households – Log(EXP) – appears to be significant predictors for both cases of probability to borrow and loan size. If the household increases expenditure by 1 percent on the average, at the 99 percent confidence level, they could have more chance to take a loan by 16 percent and get larger loan amount by 872 thousand dongs. It means the richer people have higher cash need and have larger loan size.

v     The characteristics of loan: Loan uses are important factors of loan size. The large coefficient, high significant and positive sign of variable of loan use for building house – HB – shows that if rural poor take loan for construction of house, their loan amount increases sharply by 835 thousands dongs with significant at the 1 percent level, if rural poor use loan for production, the amount of loan borrowed by rural poor rises by 329 thousands dongs with significant at the 5 percent level.

v     As expected, sign of variable – FORMAL – is positive and significant at the 1 percent level. The result confirms the conclusion in chapter 2 that size of formal loan is larger than informal loan. Combined with the result of the regression of the first-stage model, this result implies that informal credit sector mainly provides small loan to rural poor.

v     Effects of collateral on loan amount are estimated in the regression model the same as conclusion reached in chapter 2. Positive sign of coefficient implies that if rural poor household offers collateral, loan amount provided to rural poor increases by 460 thousands dongs. 

Ø      The regional location: Relative to South Central Coast (the omitted regional dummy), all other regions have positive coefficients and have statistically significant effects on probability to borrow as well as loan size except Red River Delta in the second – stage model. The results show that rural poor live in South Central Coast have least opportunity to taking loan. Meanwhile, the probability to borrow of rural poor in all six regions does not appear to differ significantly from each other at 1 percent significance level. Loan amounts provided to rural poor in Central Highlands, Southeast and Mekong River Delta tend to be much higher than that in other regions.

In summary, the findings reinforce the findings reached in chapter 2 and help us to understand deeply what factors affect most to rural poor to have high probability to take loan and obtain larger amount of loan. Important findings of the models extend our insight in borrowing by rural poor households.

CHAPTER 4 – CONCLUSIONS AND POLICY IMPLICATION

1. Conclusions

Analyzing VLSS98 data brings interesting features that improved our understanding of borrowing by rural poor in Vietnam. Some findings are:

Ø        Proportion of the rural poor borrowers and average loan sizes provided to the rural poor comprise only small part of total loans. Rural poor borrowed from both formal and informal financial sector. The rural credit market is segmented with respect to loan use. Of loan purposes, there are two major purposes including loans for production and for housing. Besides, significant sources to the rural poor for housing and consumption are relatives and friends;

Ø        Average size of loan with collateral is much larger than that of non-collateral required loans, especially average size of formal loans. It is notable that most formal loans required collateral while a large proportion of informal loans did not require collateral. Formal lenders provide cheap loans but restrict loan to production and also require collateral. The rural poor have to pay higher rate than the average interest rate. The government banks provided loans with low interest rate, some poverty programs served loans without interest rate, yet most loans without interest rate are provided by relatives and friends;

Ø        Empirical results show that characteristics of household head and household expenditure are significant predictors of probability of borrowing by rural poor and loan size provided to them. Education and social responsibility of household head have positive effects while the age negatively influence on probability of borrowing but positively on loan size. Unexpectedly, the number of household member has negative effects to probability as well as loan borrowed by rural poor. A surprising finding in the chapter is that poor women engage in borrowing much less than poor men do even though there have recently been more and more governmental non-governmental programs reserved for women; and

Ø        The final point here is the distortion of borrowing by rural poor among the regions. Under consideration of borrowing by rural poor across country, we also found that rural market is distorted among the regions. The poorer rural households are the less loan value they could borrow. Related to South Central Coast, other regions have more chance to take loan and the rural poor living in Central Highlands, Southeast and Mekong Delta took larger loan amount, especially in the Southeast.

2. Policy implications

Based on our findings about the limitation over the formal rural credit to the poor, we proposed:

Ø        The government should remove ceiling interest rates and encourage market or near-market interest rates;

Ø        If the process of authorization (land use rights) is more simply and clearly implemented, the number of rural poor who can access formal loans will increase;

Ø        Using a form of social collateral is helpful way for the rural poor who lack collateral since the group solidarity will serve as collateral;

Ø        If formal financial institutions lessen their restriction of loan to production for some other kinds of long-term investment like schooling, health care a number of rural poor can have more chance to get formal loans;

Ø        Greater redistribution to poorer and less administratively capable provinces and attention to resource distribution within provinces so that the poor are treated equally regardless of where they reside. Greater consistency in priorities and norms for identifying the poor and poor women across regions. Integration and coordination between programs with well-defined and universal rules for implementation at the local level.

In addition, the government should provide an adequate legal framework for the establishment of networks and their apex organizations for guidance, training, consultancy services, self-regulation and supervision, liquidity exchange and refinancing. In order to use their credit properly, the rural poor need training – in skill development, business, literacy, finance, agriculture and so on. That is way that the government helps the poor help themselves.