Loan Repayment Practice of Borrowers and Its Determinants: Smallholder Farmers Level Analysis

Loan is the money borrowed and must be paid back with or without interest within the time period and under the terms as agreed upon between borrower and lender. This study aimed to investigate and analyze loan repayment practice of farmers and its determinants in Amigna Gesgar Districts. To achieve the objectives of the study mainly primary data was collected from 211 respondents among 602 target population of smallholder farmers and 3 microfinance institutions officials of the district were included and some extent secondary data was incorporated; in analyzing the collected data descriptive and econometric approach were used. The results of the study shows that about 30.34% and 69.66% sample borrowers were defaulters and non-defaulters respectively.  Additionally, the binary logit regression model result reveals, the purpose of the loan, experience the farmers, dependency ratio, and wealth the farmers were the determinants that affect the probability of loan repayment significantly at 5% probability level. Finally, it was strongly recommended that, the smallholder farmers should use the money borrowed only for productive economic activities in order to repay the loan on time and the government and private financial institutions should improve services of credit and saving institutions at the rural area of the study area. Keywords: Dependency Ratio, Loan Repayment, purpose of loan, Smallholder Farmers DOI: 10.7176/JESD/11-11-05 Publication date: June 30th 2020

farmers and its determinants at district. The research the main research questions were:  What are customer's farmer's capacity in repayment of the loan in Amigna Gesgar District?  What are the determinants of loan repayment practice of farmers in at the study area?  How the problem of loan repayment can be reduced? Significance the study The study will have numerous importance, among those the following are the potential significances. It was be important to create awareness in borrowing loan from microfinance by smallholder farmers. Second, it identifies the major determinants, depend on the identified farmers will reduce loan repayment on time. Third, microfinance institutions may change their policies, strategies to help poorer smallholder farmers depend on the identified outcomes. Moreover, this study which identified and raised some main points related to the loan repayment practice of farmers and its determinants issues and there by stimulates further investigations.

II. Review of literature 1. Theoretical Literature Review i. Concepts of Micro Finance Institutions and Credit Policy
According to Wood (1998) cited in Walter (2013, defined micro finance institutions as institutions that provide financial services to low income earners. These financial services may include savings, loans, insurance transfer and payment services to enhance growth of small scale enterprises. According to Graham et al (1990) microfinance. refers to the provision of financial services to the low-income earners who do not earn or obtain their services from the formal financial institutions because of their business saving levels and credit needs are very small. To Bank of definition, microfinance institutions are non-governmental institutions savings and credit co-operatives that provide savings and microloans not exceeding US $ 5000 for project to poor individual's average not exceeding US $1000 per project to poor individual enterprises or groups for the purpose of engaging in income generating activities.
A credit policy is an institutional method for analyzing credit requests and its decision criteria for accepting or rejecting applications. Credit policy is important in the management of accounts receivables. A firm has time flexibility of shaping credit policy within the confines of its practices. It is therefore a means of reducing high default risk implying that the firm should be discretionary in granting loans.
Policies save time by ensuring that the same issue is not discussed over and over again each time a decision is to be made. This ensures that decisions are consistent and fair and that people in the same circumstance get treated in the same manner (Khandkar and Khan, 1998). A credit policy provides a frame work for the entire management practices. This provides risk protections by enabling the lending institutions to follow up when the borrowers fail to honor the agreement. Microfinance is a form of financial development that has primarily focused on alleviating poverty through providing financial services to the lower income society or the poor. People think of microfinance, if at all, as being about microcredit, which is lending small amounts of money to the poor. Microfinance is not only offering this service only, but also it provides broader: including insurance, transactional services, and savings (Barr, 2004).
The characteristics of microfinance products include (Murray and Boros, 2002, pp.10-11): small amounts of loans and savings, short-period loans (up to one year term), payment schedules attribute frequent installments or frequent deposits, payments (installments) made up from both principal and interest, higher interest rates on credit (higher than commercial bank rates), easy access to the microfinance intermediary saves the time and money of the client and permits the intermediary to have a better idea about the clients' financial and social status, simple application procedures, short processing periods (between the completion of the application and the disbursement of the loan), the clients who pay on time become eligible for repeat loans with higher amounts. No collateral is required contrary to formal banking practices. instead of collateral, microfinance intermediaries use alternative methods, like, the assessments of clients' repayment potential by running cash flow analyses, which is based on the stream of cash flows, generated by the activities for which loans are taken.

ii. Microfinance in Ethiopia
Microfinance institutions introduced in Ethiopia after the dawn fall of the Derg regime following the policy of economic liberalization. The development of microfinance institutions in Ethiopia is a recent phenomenon. Ethiopian development strategy is the establishment of sustainable microfinance institutions serving large number of poor people. Microfinance is taken as a shift from government and NGO subsidized loan programs to finance services run by specialized financial institutions. Later microcredit programs were changed to microfinance institutions. Non-governmental organization (NGO) credit schemes and informal sources of finance have existed in Ethiopia for many years; the government instituted a legal and policy framework for MFIs in 1996 through Proclamation 40/1996 (Gebrehiwot, 2002).
Currently, the Ethiopian microfinance industry is rapidly growing. Recently, there are 31 MFIs reaching around 2.4 million people (AfDB, 2012). The deposit interest rate is 3-8% and lending interest rate is 12-24% in 2012. Most MFIs are doing remarkably well in terms of financial performance given their relatively short track Journal of Economics and Sustainable Development www.iiste.org ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.11, No.11, 2020 59 record (Facet, 2013). Fikirte (2011) on her study loan repayment performance of Addis credit and saving institution (AdCSI) in Addis Ababa, Ethiopia she employed a binary logit model used to analyze socio economic factors that influence loan repayment. The study shows out of 11 variables age and five business types (baltina & petty market, kiosk & shop, services providing, weaving & tailoring and urban agriculture) were important in influencing loan repayment performance of the borrower. In this study loan provision in group based lending in terms of the dependency ratio was negatively and significantly related to being defaulter, implies that borrower family in the group lending might be involved in productive activities. Furthermore, sex and business experience of the respondents were found to be insignificant and positively related to loan repayment rate. Characteristics of Microfinance gives access of financial services to low-income people, who wish to access money for starting or raising an income generation activity. The individual loans and savings of the poor clients are small. Microfinance came into being from the appreciation that micro-entrepreneurs and some poorer clients cannot have access to borrow from banks, that is, they can repay, both the principal and interest, on time and also make savings, provided financial services are tailored to suit their needs. Microfinance as an institution has created financial products and services that together have enabled low-income people to become clients of a banking intermediary.

Empirical literature
The existing theoretical models of peer monitoring deduce that repayment performance in group lending programs is positively related to the homogeneity of members with respect to the riskiness of their projects. In group-lending programs, the functions of screening, monitoring and the enforcement of repayment are to a large extent, transferred from bank to group members. The financial intermediary reduces recurrent lending transaction costs by replacing a multiple of small loans to individuals with a larger group loan (Zellar, 1998). This reduction in transaction costs enables financial intermediaries to bank with poor, who demand small loans and who would not receive any credit under an individual loan contract because of excessive unit transaction costs of tiny loans.
It was argues that probably the most important rationale for group lending is the information and monitoring advantages that group-based financial institutions at the community level have, compared to individual contracts between a bank and borrower. Group members get important information like reputation, indebtedness and asset ownership of the loan applicants at a lower cost (ibid). They can also easily monitor individual efforts made towards ensuring repayment. On the other hand, groups may also have a comparative advantage in enforcement of loan repayment. Moreover, group members appear to be in a better position to assess the reason for default and to offer insurance services to members who are experiencing shock that are beyond their control.
Individuals select those whom they thrust to form a group with that is they want those who can make regular repayments, have a good concern about the possible loss they face in case of non-repayment, ultimately leading to the exclusion of the poorest of the poor. According to him presence of high geographical mobility, low attachment to specific neighborhoods and peer groups consisting of competitors are the factors that frustrate the solidarity of groups in urban areas, and hence group lending is more applicable to the rural environment than to urban society.
Million et al., (2012) studied on factors affecting loan repayment performance of Small holder farmer in Eastern Harergie, Ethiopia. In order to analyze the factors that influenced loan repayment performance, employed a two limit tobit regression model and also structural questioner was used to collect information from 140respondent in two district (Kombolcha and Babile ),by using multi stage sampling technique. The result shows that agro ecological zone, off farm activities and technical assistance from extension agents positively influence loan repayment performance, while product loss, informal credit, social festival are negatively related to loan repayment. Abafita (2003) analyzed the microfinance repayment performance of Oromia Credit and Saving Institution in Kuyu, Ethiopia. According to his finding; sex, loan size and number of dependents are negatively related to loan repayment. On the other hand age was found to be positive, while age squared turned to be negative. Income from activities financed by loan, repayment period suitability and loan supervision are positively and significantly related to loan repayment performance. Moreover, loan diversion is significant and negatively related to loan repayment rate. The negative sign implies that the use of diverted funds for non-income generating purposes. Zeller (1998) analyzed the determinants of repayment performance of credit groups in Madagascar. His finding is groups with higher level of social cohesion have a better repayment rate. Moreover, the programs that provide saving service to their members have a significantly higher repayment rate. Olagunju & Adeyemo (2007) and Oke et.al. (2007) also analyzed the determinants of repayment decision among small holder farmers in southwestern Nigeria. The result showed that the number of visits made by loan officers to the borrowers, higher level of education, and time of loan disbursement would have a better repayment performance.

III. Aim and Objective of Study
The general objective of the study was to assess the loan repayment practice of farmers and its determinants in case of Amigna Gesgar District, Arsi Zone, Oromia National Regional State, Ethiopia. And the specific objective of the study were the followings:  To assess the capacity of farmers in the repayment of loan at the study area.
Journal of Economics and Sustainable Development www.iiste.org ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.11, No.11, 2020  To examine how to reduce the problems of loan repayment performance of rural households  To identify the factors that affect farmers in loan repayment at the study area. Hypothesis of the study Several statements of supposition can be made in view of loan repayment practice of smallholder farmers and its determinants at the study area. The following lists of hypotheses 1 are the major ones on which the study is pivoting. Hypothesis 1: The dependency ratio has no significant impact on loan repayment of smallholder farmers at the study area. Hypothesis 2: The purpose of money borrowed by a smallholder farmer and loan repayment has no relationship. Hypothesis 3: Education of smallholder farmers has significant impact on the probability of loan repayment of at the decreases. Hypothesis 4: Education of smallholder farmers positively affect the probability of loan repayment of at the study area. . Hypothesis 5: The level of wealth of smallholder farmers has insignificant impact on the probability loan repayment of the smallholder farmers at the study area. Hypothesis 6: Marital status of smallholder farmers has significant on his/her loan repayment of at the study area.

IV. Research Methodology 1. Description of the Study Area
The study was conducted in the Amigna Gesgar, one of the districts of Arsi zone, located in Oromia Regional State. Amigna is one of the districts of Arsi zone. According to the information obtained from elder persons of the area, the name Amigna was derived from one of the Oromo sub-clans, which is known as Amigna. It was recognized as an administrative district in 1958E.C. The district shares boundary line with West Hararghe Zone in the East, Gololcha and Chole districts in the north, Sude and Robe districts in the west and Belegesgar, Seru district in the south. Regarding aerial coverage, Amigna district has a total area of 1314 km accounting for about 5.6% of the total area of the Arsi Zone. Adelle town is the capital of the district; it is located at 254km from Addis Ababa/Finfinnee capital city of Ethiopia and 129.71km from zonal capital, Asela. (OBFEC, 2019).

Sampling Technique and Sample Size
The researcher used the mix of purposive and simple random probability sampling techniques; first the four sample gandas 2 was selected purposively. As the Central Statistical Agency of Ethiopia (2007), the size of estimated population of Amigna Gesgar was totally 73,245; among this 91.23% (6,820) were rural residents. According to this data, from the total residents of the rural area of district 33, 294 and 33, 523 were males and females, respectively. By considering this data, the researcher estimates the population size of the selected four gandas at the period of the data collection and estimates the number of their households head as 602; this was the size of target population which includes both defaulters and non-defaulters, smallholder farmers. Then, sample size was determined by using 95% level of confidence, 9% degree of variability and margin of error (e) (Yemane, 1967cited in Bahabelom, 2010. n= Where, n: sample size, N: population size e: level of accuracy A total of 211 respondents were taken from the population size of 602 smallholder farmers of the four selected small administration units of the district.

Data Source and Method of Data Collection
The researcher has been used both primary and secondary data; mainly primary data source. The data was collected through questionnaire, interview and focus group discussion in order to get more information in detail about the title. The target respondents were 211smallholder farmers, and 3 officials from the district. And the secondary data was collected from different both published and unpublished documents that concerns with the selected title.

Research Design
Research design is a backbone of a paper in specifying the method and procedures for collecting and analyzing the information. The quantitative approach involves the generation of data in quantitative form which can be subjected to rigorous quantitative analysis in a formal and rigid fashion. However, qualitative approach to research is concerned with subjective assessment of attitudes, opinions and behavior (Kothari, 2004). This study is analytical to establish the degree of relationships between some relevant factors and issues as well as to show the relative size or significance of each factor relative to the others. The types of research employed under this study was be descriptive and econometrics approach. The major purpose of descriptive research is description of the state of affairs as it exists at present. 1 All of the listed hypothesis were null hypothesis while the opposite idea of each null hypothesis was alternative hypothesis. 2 ganda -refers to the lowest level of administration in a given district or town; and gandas or gandoota-plural form of ganda Journal of Economics and Sustainable Development www.iiste.org ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.11, No.11, 2020

Method of Data Analysis
Data collected data was analyzed by using both descriptive method and econometrics approach. In the descriptive analysis, the researcher used descriptive statistics. In the second part econometric issues, more specifically, Logit model was used adopt. Also, the independent variables those significant and insignificant independent variables would be as identified. Additionally, the correlation among independent variables was determined by using recent STATA software.
6. Description of Variables A. Dependent Variable: in this research design the dependent is the probability of loan repayment by smallholder farmers; this implies that the dependent variable is dummy variable. B. Independent Variables a. Age (X1): it refers to the age of borrowers, smallholder farmers and measured in years. Thus, it was continuous variable. It was expected to affect the loan repayment negatively, due to as the age of the borrower increases his production and productivity will decrease; as a result his/her income decreases. Therefore, as level of the borrower's age increases loan repayment decrease due to their ability of work decrease. b. Marital status of farmers or the borrower (X2): It refers to a situation that a borrower, smallholder farmer was being married or single. It was dummy variables, if the farmer has married/non-single, it was take 1 otherwise if he/she single (which includes unmarried, divorce and widow) it was take 0 value. It is expected that if the borrower have husband/ wife doesn't repay the loan on time due to after marriage, the number of children increases and dependency ratio become high. c. Education level of farmers or the borrower (X3): It was taken as continuous variable in terms of years of schooling. It was expected that there was positive relationship between loan repayment and education level of the borrower, because as a borrower level of education increases, it was supposed that his/her income also rise. d. Purpose of the Loan (X4); it refers to the purpose of money borrowed and it was it is a dummy variable taking 0 for loans used investment activities and 0 if it used for consumption purposes of for households. It is argued that loans used for investment rather consumption purpose have positive impact on successful loan repayment performance of the borrowers. e. Experience (X5): it refers to the experience of the borrower, smallholder farmers in business; and it was a continuous variable which was measured in years. Borrowers who have been in business longer are expected to be more successful with their business. f. Ratio of dependency (X6): it refers to the ratio of the number of children whose age were below 15 years and old man/woman whose their age is greater than 65 years to the number labor force in the household of the borrowers; and it was continuous variable. g. Wealth (X7): it refers to the wealth 1 level of the borrower and measure in birr; it was continuous variables.
The amount that households spend and save from current income depends partly on the value of the existing wealth they have already accumulated. And it was expected that higher level of wealth of the borrower the probability of loan repayment would be also high.

The Model Specification
The dependent variable was the probability of loan repayment by borrower, smallholder farmer's performance; has two outcomes: either pay or not pay the borrowed money. Arbitrary values 1 and 0 can be assigning for pay or not pay respectively. For a dummy dependent variable, binary responses models such as logit and probit models are best suit for analysis. As Hosmer and Lemeshew (1989) pointed out, a logistic distribution (logit) has advantages over the other in the analysis of dichotomous outcomes variable in that it is an extremely flexible and easily usable model from mathematical point of view and results in a meaningful interpretation .The researchers was employee the logit model for its simplicity in this research (Gujarat, 2009). In short, the dependent variable of a logit model takes a binary response. That means Yi= 1 if a borrower was pay and 0 if Yi= 0 if a borrower was not pay. The form of the Logit model is shown as follows. Y=α + β1 X1 + β2 X2 + β3 X3 +…+ βkX k+ԑ …………………………………………… 1 Where, Y = Probability of a loan repayment α = Intercept (constant) term βi =Coefficients of the predictors estimated using the maximum likelihood method Xi= Predictors (independent variables) ԑ = Random effect (error term) Let's suppose that the response variable * y, captures a true status of the borrower either as pay or non-pay so we can estimate the regression equation as follows is not observable and is a latent variable. Aggregating the value yields. *y = α +∑ β + ԑ ………………………………………………………………………… 2 In practice *y is unobserved, and ԑ is symmetrically distributed with zero mean and has cumulative distribution function (CDF) defined as F (ԑ). What we observe is a dummy variable y, a realization of a binomial process defined by y= 1, ˃0 0 ℎ ! " ……………………………………………………………………. 3 From equation (1) leaving the constant term and rewriting the model yields The logit model usually takes two forms. It may be expressed in terms of logit or in terms of event probability. When expressed in logit form, the model is specified as Log# Using equation '4' and '5' can be transformed into a specification of the logit model of event probability by replacing the general CDF, F, with a specific CDF, L representing the Logistic distribution ………………………………………..………………………………… 6 The above equation was represent the probability of an event occurring. For a non-event, the probability is just 1 minus the event probability.
Also since the above equation zi= β 0+ β 1 x 1 + β 2 x 2 + β 3 x 3 +...+ β i x i + Ei from the above model let us substitute zi with dp(dummy of participation) as the dependent variable.
Where dp= dummy of poverty, Ei is error term, β 0 is constant and β's are coefficients. The researcher specify y the model as follow through the relationship between dependent and independent variables. Y=f (X1, X2, X3, X4, X5, X6, X7) Where: Y=borrower's loan repayment status, X=independent variables that explained above. Addition to the above points, to have relevant output particularly in the econometrics model the necessary test of the binary logit regression model which includes correlation test, good fitness of the model test were employed.

V. Data analysis and Interpretations 1. Descriptive Analysis
The collected data was analyzed and interpreted by Table's frequencies and percentages was used to identify the main determinants of loan repayment of the borrowers, smallholder farmers at the study area.
A. . Demographic Characteristics of the Respondents. i. Age and sex of Respondent As shown in the following table the sex of the respondents who participate in responding, or filling the questionnaire were both males and females. But, in interview and focus group discussion all of them were females. The above Table 1 shows that the respondents includes both males and female, 90% of the respondents were males and the left were females. Among the selected respondents 28.44% were defaulters during the data was collected. From this result, it can be concluded that the sex of the respondent has insignificant association with the loan repayment at the study area. On the other hand, as explicated in the Table 2 below, from the total selected respondents 30.34%, 69.66% were defaulters and non-defaulters respectively. The result shows the number of both defaulters and non-defaulters of the selected respondents, 54 and 179 found in labor productive age. Even though Journal of Economics and Sustainable Development www.iiste.org ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.11, No.11, 2020 from the selected respondents 0.95% of them totally from the age of above 64(from their age group 100%), it can be summarizes as the relationship among the age of borrowers (smallholder farmers) and the probability of loan repayment is weak. Also this fact was explained econometrics approach, binary logit regression model (see Table  4 below). Marital status of the respondents Marital status of the respondents was another interest of the researcher in the sense that it explains the extent of one's responsibility, therefore it was investigated and the following Table shows the response.
According to the Table 3, 26% among the defaulters of the respondents which is the biggest percentage were married while 3.3% and 1.9% of the respondents single and divorced respectively. As the collected data indicates that, from the small holder farmers at the study area about 76.7% of them were married while 47 of respondents (22.3 %) were single (unmarried) farmer, borrower. However, as explained in the table among defaulters, most of them 55 (86% from the total number of defaulters) or 26% from the total respondents were married; this data reflects that these households who borrow money in the marital status of married negatively affect the probability of loan repayment at the study area.

. Econometric Analysis of the Determinants of Loan Repayment
An economic model, logistic regression model, binary logistic model was used, by independent variables used in order to identify determinants of loan repayment of smallholder farmers at the study area. Binomial (or binary) logistic regression is a form of regression which is used when the dependent is a dichotomy and the independents are of any type. In this research, the dependent variable, that is. the status of the probability of loan repayment on time fully, which implies it is dichotomy; while the in independent variables include both dummy and continuous variables as explained above. Therefore, the best model is binomial (or binary) logistic regression (Gujarati and Porter, 2009).
The result of binary logit regression model with coefficients helps simply to understand the direction of relation between the probability of loan repayment and each independent variable like what happen in OLS, even though their interpretation in magnitude is different and the result was the followings:  Vol.11, No.11, 2020 Additionally, we can calculate t-statistics to in order to check the significances of each independent variable. For instance, it can be calculated for the purpose of the loan in order to check its significances status in the binary logistic model as below: * = 4 1 &5 1 6 (4 1 ) = 4 1 6 (4 1 ) = & . 78 9: . ; 9

= │-2.13│ = 2.13
The value of the t-statistic that corresponds to the 5 % (commonly used) significance level is 1.96 and 2.13> 1.96, so we reject the null hypothesis. That means the estimate is significantly different from zero at 5% significant level. The same is true for others independent variables also. As explained in the Table above, seven independent variables are hypothesized to have influence on the probability of loan repayment of borrowers included in the model of which s four of them were found to be statistically significant and the sign of most independent variables were as expected; purpose of the loan, experience the smallholder farmer, dependency ratio of households of smallholder farmer or the borrower and wealth the smallholder farmer.

Interpretation of Odds Ratio in the model
In this study, dependent denoted as the probability of loan repayment bay smallholder farmers that determined by various independent variables. In the other words, odds of the model tell by what factor the dependent variable change does whenever a unit change occurs in an independent variable (Greene, 2012). OR 1 is the Log value of odds and is always positive. In this particular study it is found out that the odds ratio, the ratio of the probability of being participant to the probability of non-non participant. The detail let's interpretation logit model with odd ratio was as the followings: Age of the smallholder farmer/the borrower (age_brwr) In the study, the age of the borrower was measured as continuous variable education level. The result of the study shows that as the age of the borrower increases the ability/ performance to repay the money borrowed on time increases because as the age of the borrower increases his/her income from farm and non-farm economic activities increases. And, as explained in the Table 5 above, even though in the result of the model the age of the borrower (smallholder farmer household head) had insignificant even at 10% probability level, i.e. it had less impact on the dependent variable. ii.
Educational Level of the smallholder farmer/the borrower (educ_brwr) In the designing of the research, the education level of the borrower taken as continuous variable and measured in the number of years the stay in school. In the model of the study, as demonstrated in the Table 5 above, marital status of smallholder farmer/ the borrower was insignificant; that is , it had less impact in determining the performance farmers (as borrowers) in loan repayment on time at the study area (rural). iii.

Marital Status of the smallholder farmer/the borrower (maritalstu_brwr):
In the model of the study, marital status of smallholder farmer, household head at the study area was insignificant variable at 5% probability level. iv.
Purpose of the Loan (purpose_loan) The model result, at the study area the purpose of the loan was significant at 5% probability level (as explained in the Table 5 above). And the interpretation of the result is as a smallholder farmer (the borrower) increases the money borrowed for investment purpose by one percent (going from 0 to 1), the probability of repayment of loan increases by 6.92 times greater, given that all of the other variables in the model are held constant.

v.
Experience the smallholder farmer/the borrower (bexper_brwr) 1 Odds Ratio -The odds ratio is a measure, or it approximates how much more likely (or unlikely); it is for the outcome to be present among those with independent dummy variables is equal to one than among those independent. Odd ratio in favor of participation in economic activities -the ratio of the probability that a woman participate in economic activities to the probability that she will not participate in economic activities (Green. 2012).  Vol.11, No.11, 2020 As demonstrated in Table 5, the experience the smallholder farmer on utilizing loan on business activities has significant impact on the dependent at 5% probability level. And the interpretation of the result is as the experience on utilizing loan for business investment of smallholder farmer (the borrower) increases by one year, the probability of repayment of loan increases by 2.27 times greater, given that all of the other variables in the model are held constant. vi.
Dependency ratio of households of smallholder farmer or the borrower (dependency_ratio) As presented in the Table 5, the dependency ratio of households of smallholder farmer has significant impact on the dependent variable, the probability of loan repayment successfully at 5% probability level. And the interpretation of the result is as the experience on utilizing loan for business investment of smallholder farmer (the borrower) increases by one year, the probability of repayment of loan increases by 0.108 times less, given that all of the other variables in the model are held constant. On other hand, borrowers with lower number of children household members (lower dependency ratio) would meet their repayment obligation better than those with high number of children household members. vii.
Wealth the smallholder farmer (wlth_brwr): As explained in the Table 5, the wealth the smallholder farmer of households of smallholder farmer has significant impact on the dependent variable, the probability of loan repayment successfully at 5% probability level. And the interpretation of the result is as the wealth of smallholder farmer (the borrower) increases by one percent, the probability of repayment of loan increases by 1.01 times greater, given that all of the other variables in the model hold constant. To the extent that one independent is a linear function of another independent, the problem of multicollinearity will occur in logistic regression, as it does in ordinary linear regression (Green, 2012). Thus, the correlation among continuous independent variables was also tested as presented in the Table 6 above; and the result shows that, the correlation results demonstrates among continuous variables there was weak correlation, for the most of independent variables the values were less than 0.4, which implies the problem of multi-collinearity didn't occur in the binary logit regression model. The marginal effect of the model was computed as explained in Table 7 above. For instance, the result of marginal effects shows, as a smallholder farmer has a one year in business experience (as borrower) has a 2% higher probability of loan repayment.
Addition to the independent variables used in the model above, the as determinants of smallholder farmers, the result of the collected data through interview and focus group discussion demonstrates others factors also affect the probability of loan repayment. Among them, first expectation of the future can influence the savings of smallholder farmer, because if prices of goods and services consumed by a farmer expected to fall in the future, their present consumption is less, and hence saving is more. In this case the probability of loan repayment fully on time will high. Secondly, extra expenditure during different social activities like annual celebration, and wedding. And in each activities there was competition among each households of the farmers in which extravagancy was mentioned. The other was the nature of the smallholder farmers: in nature some smallholder  Vol.11, No.11, 2020 farmers were save more while the others save less. Such, borrowers have high probability in the loan repayment at the study area.

VI. Conclusion, Findings and Suggestions Conclusion and Findings
This study aimed to investigate and analyze loan repayment practice of Farmers and its Determinants in Amigna Gesgar Districts, Arsi Zone, Oromia Regional State, Ethiopia by employing cross sectional data which was collected in 2010/11 E.C. To achieve the objectives of the study mainly primary data was collected from 211 sample size from four selected gandas of the district. In the analysis of the collected data both descriptive and econometrics model was used. This study were conducted to identify socio-economic and institutional factors affecting loan repayment. In order to improve and enhance the life of smallholder farmers and rural economy, credit and saving institution is very important because of: poor households are too poor to save, poor to bridge consumption, thus unable to help them; poor farmers lacked access to formal credit services; the supply of institutional credit is an essential requirement for (agricultural) development regardless of demands; its credit should be cheap (subsidized) due to smallholder farmers are poor to afford market interest rates.
The result of the study demonstrates that, the general determinants and constraints of loan repayment were age, education level, purpose of the loan, experience on business, dependency ratio, wealth level, expectation on price of goods and services, marital status, and the nature the smallholder farmer or borrower; and the expenditure level on social festivals like during annual holidays, weddings, and others social factors were identified identifies as the major determinants in the study. Additionally, the of binary logit regression model result shows that the wealth of the farmer (which includes both farm and non-farm income of the farmer), the purpose of the money borrowed, the experience of the farmer (borrower) in business, and dependency ratio were significant at 5% probability level in influencing the dependent variable, the probability of repayment of the loan, the money borrowed by the farmer. The descriptive statistics results showed that about 30.34% and 69.66% percent of sample borrowers were defaulters and non-defaulters respectively.
The purpose of the money borrowed affected loan repayment performance positively and significantly. This is due to the fact that those borrowers who participate in business activities rather for consumption purposes, their probability to pay the loan was high, because they were earn more income from the business or investment(s). As a result explained in econometric model of the study, as a smallholder farmer (the borrower) increases the money borrowed for investment purpose by one unit (going from 0 to 1), the probability of repayment of loan increases by 6.92 times greater, given that all of the other variables in the model are held constant.
The second explanatory variable which has a significant impact on the dependent variable in the model was the dependency ratio of households of smallholder farmer (borrower); it affected loan repayment performance negatively and significantly. Here the collected data shows that those households borrows who's their number of children was large the dependency ratio become; and as a result the probability of the loan repayment become low. In short according to the result of the model, the probability of repayment of loan increases by 0.108 times less, given that all of the other variables in the model are held constant. The third explanatory variable which was significant in the model was the experience the smallholder farmer/the borrowers. The collected data shows that those households borrows that as the experience of the farmer (borrower)in business activities the probability of repayment of loan increases by 2.27 times greater, given that all of the other variables in the model are held constant. The final independent variable that was positively and significantly affect the independent variable, the probability of loan repayment was wealth the smallholder farmer of households (as borrower). And the result shows that as the wealth of smallholder farmer (the borrower) increases by one percent, the probability of repayment of loan increases by 1.01 times greater, given that all of the other variables in the model are held constant.

Suggestions
Based on the findings of this study, the researcher found it important to make some recommendations to guide the enterprises, other concerned bodies and researches.  Microfinance officials, and smallholder farmers (borrowers) should be trained by professionals how to develop business plan. The culture of cooperation, and formal and informal should be improved by taking the work of successful enterprise repayment period should be revised, because to be profitable and repay the loan easily adequate repayment period is required.  Additionally, there was a problem of smallholder farmers' attitude toward business activities in general; thus, government (concerned body) should deeply create awareness on how to create business and success in the business process for smallholder farmers (borrowers) at the study area.  The smallholder farmers of the study are should minimize extra expenditure during different social activities like wedding, and annual festivals.  The members of farmers' households should stay single until they have sufficient income for their family to live independently; due the result of the study shows that single/unmarried farmer save more than those nonsingle/married and their probability of loan repayment was high as explain indirectly under dependency ratio.  The government should construct the concrete asphalt which connects to Adele Town to Adama/Assela to support small farmers in providing their products in supplying at center market in order to get the balanced price of their outputs; as a result their income increases and the probability of loan repayment will be high.