The Impact of Microfinance on Multidimensional Poverty Status of Rural Households in Gozamen District, East Gojjam Zone, Ethiopia

Microfinance aimed at breaking the vicious circle of poverty in Ethiopia mainly by providing loan service for rural households. The main objective of this study is to examine the impact of microfinance loan service on multidimensional poverty status of rural households by taking evidence from Amhara Credit and Saving Association. To attain this objective, the researchers collect primary data by using household survey from the total of 290 sample sizes 145 from treated group and 145 from non treated group respondents by using quasi experimental design. To analyze the data, the resrahcers employed descriptive statistics and inferential statics. The propensity score matching model result reveled that microfinance loan service has a negatively impact on the multidimensional poverty status of rural households. It is also found that microfinance loan service has reduced standard of living, health and educational dimensions of poverty respectively for rural households of the study area. It is recommended that government should give special attention to support microfinance’s who support the rural poor household heads and improve the awareness level of farmers about its role towards poverty reduction . The data shows that standard of living take the largest domain of multidimensional poverty status of rural households in the sampled area of Gozamen district .The finding of the study confirm with (Andualem, 2016).

are female. The total number of households accounts about 30,146. It has a total area of 1281,065,863 with a population density of around 119 per square kilometer. Gozamen district has different landscapes, most of which are mountainous. The altitude ranges from 800m-2400m above sea level. This makes the district to have kola, Woyina Dega and Dega climatic regions. The majority of the populations' economic activity depends on agriculture. Around 97% of the population is dependent on agriculture. Only few have additional source of income from weaving, poetry and small business. Gozamen district is dominantly a food crop producing area of which Teff, Wheat and Maize are the most common outputs (Amhara bureau of finance and economic development, 2012).
The study followed quantitative research approach since the nature data for this study is quantitative nature. This study also investigates the poverty status of rural households and the extent of microfinance loan service on poverty status of rural household by taking evidence from Amhara credit and saving association and hence, the researchers follow descriptive research design.

Data Types, Sources and Methods of Data Collection
The study employed primary data and the method of data collection was household survey collected by structured questioners.

Sampling Techniques and Sample Size Determination
The population refers of the study were all household who are registered as a user of Amhara Credit and Saving Association in Gozamen district. The populations at which the samples were drawn are mostly located in rural areas. Probability sampling technique was used in the process of data collection. The population of Gozamen wereda is homogenous in many aspects except agro-ecological difference. Gozamen wereda has three climatic regions; Dega,Woina Dega and Kola agro-ecological zones. Based on this difference a stratified sampling technique was used to group the sample kebeles. There are 5 Dega, 2 Kola and 18 Woyna Dega kebeles.
A total of four kebeles were selected from the total kebeles of 25 using simple random sampling method, randomly drawn from a complete list of kebeles. One kebele each was selected from Dega and Kola areas while two kebeles were selected from the Woyna Dega area with the principle of proportional representation. A complete list of microfinance users and non-users was collected in each kebele and a proportional sampling was taken from both users and non users. The sample size was determined in proportion with the agro-ecological zones and the number of microfinance services user and non-user households.
Once the sample kebeles are identified, a sampling frame which contains a complete list of households (3202) was prepared and the sample determined using a simple formula (Cochran 1977). no= P is the estimated proportion of an attribute that is present in population which is incidence of poverty. q is 1-P, e is significance level (5%) ,Z is standard normal distribution ( z 2 =3.8146) and n is sample size. According to Tsegaye (2014) incidence of poverty in Gozamen wereda (p) is 0.31 and q will be 0.69.
Based on this, we have got 327 households. But when the sample size is more than 5% of the sample fame, Cochran (1977) suggested correction mechanism as; n= / N is sample frame and no is sample size in the original equation. By the correction mechanism, we have got 290 households. The sample for each kebele is obtained by using; nk = * 290 and the kebele sample is divided between microfinance service users and non-users in the same procedure.
The sampling procedure and sample size is seen in the following table.

Methods of Data Analysis and presentation
The study used both descriptive statistics, and inferential statistics. The descriptive statistic was summarized using average, and percentage to show the multidimensional poverty index, head count and intensity poverty. The data is presented in the form of table. For inferential statistics, propensity score matching model is used to examine the impact of Amhara credit and saving Association loan service on multidimensional poverty status of beneficial rural households compare to that non beneficiaries in Gozamen district since the microfinance loan service were non randomly assigned for users.

Variable Selection and Model Specification 2.5.1 Variable selection
To measure the multidimensional poverty status the researcher use three dimensions of poverty such as education, health and standard of livings and ten indicators of poverty as listed below ; Depended variable Head count of Mutidimestion Poverty (H):-multidimensional poverty head count status of each household as dummy dependent variable or outcome variable. It can be labeled 1 for poor 1, other wise 0. The cut off head count is determined by when MPI equal to 0.33 and above the household considered as poor and other wise non -poor adopted from (OPHI, 2017 The study uses ten indicators of deprivation as follows; Adult education deprivation: -Education Indicator-Years of Schooling, dummy variable (0=ND,1=D).

Child education deprivation:-Education
Indicator-School attendance, dummy variable (0=ND,1=D). Nutritional deprivation:-Health Indicator-Adult malnutrition (0 =ND, 1=D), nutritional status is taken from the computation by using direct calorie intake of households. If the household takes less than the standard per capital nutritional requirement 2,100 calorie per adult per day set by the Ethiopian government the household is deprived(D), otherwise non-deprived(ND). Child mortality: -Health Indicator -Child Mortality, (0=ND,1=D). Floor Derivation:-Standard of Living Indicator -Flooring or roof dirty material like grass (0=ND,1=D) Sanitation deprivation:-Standard of Living Indicator-improved sanitation (0 =ND,1=D). Access for clean water deprivation:-Standard of Living Indicator-Access for clean water (0 =ND, 1=D).Given that less than 30 minute walk fetch and come to home. Energy deprivation:-Standard of Living Indicator-Cooking Fuel (0 =ND, 1=D) Elcctric city deprivation:-Standard of Living Indicator-Electricity (0 =ND,1=D). Asset deprivation:-Standard of Living Indicator-Assets (0=ND, 1=D). Asset deprivation represents absence of least the following assets such as; television, Animal cart, and bicycle and farming tools The weight of the above three dimensions and ten indicators will be adopted from OPHI 2017. The methodology of computing MPI can be done as follows; 1. To choose the poverty deprivation cut off (identify which household is poor). Each person is assigned a deprivation score according to his or her deprivation in the component indicators which lie between 1 and 0. It can be expressed as; = + + + , where I=1, if the person is deprived in indicator "i", and I=0, otherwise and is the weight attached to indicator "i" with sum of weight equal to 1. With any combination of the indicators any one will be multidimensional poor if and only if; MPI is greater than or equal to 0.33, multidimensional poor. 2. Computing the MPI (aggregation).
i. Calculate the multidimensional poverty Head count (H): the percentage of people who are poor which shows the incidence of poverty. It can be expressed using the formula: = where "q" is the number of people who are multidimensional poor and "n" is total population. ii.
Calculating the Intensity or Breadth of poverty (A):-It is the average deprivation score of multidimensional poor people or the average percentage of dimension in which the poor people are deprived. It can be expressed as; , Where ci(K) is censored (for those whose deprivation score is below poverty cut off , even it is non-zero this is replaced by zero) deprivation score of individual ( i), and q is the number of people who are multidimensional poor . 3. The Calculated multidimensional poverty index (MPI) for the study area measures the proportion of weighted that the poor experience in a society out of all the total deprivation that the society could experience. The MPI can also be broken down by indicators, which is a useful tool for public policy. It means that MPI itself is simply the percentage of people who are poor and deprived in each indicator multiplied by the weight on that indicator. it can be expressed as; MPI =H×A, where "H" is head count ratio, and "A" is intensity ((OPHI), 2017).A person identified as poor if he /she is deprived in at least one third (33.33 percent) of the weighted indicators ((OPHI), 2017).

Model specification
The dependent variable is a dummy that takes a value of 1 when a household is multidimensional poor and 0 otherwise by using 0.33 as a cut off adopted from OPHI, 2017. To examine the impact of Amahara Credit and Saving Association loan service on multidimensional poverty status of rural households the propensity score matching model estimated with logit model is used.
Estimating the average treatment effect can be as follows; 1 N1 is number of participants and No is number of nonparticipants i index of participants and j index of nonparticipants Wij weights.

Descriptive statistics 3.1.1 Multidimensional poverty status of Rural Households in Gozamen District
This study found that 73.81 percent of rural peoples in Gozamen district are multidimensional poor, on average the poor people are deprived in 49.18 percents of the weighted indicators and the society is deprived in 36.30 percent of the total potential deprivation it could experience over all. Rural households in Gozamen district are deprived at least either all indicator of a single dimensions or a combination across dimensions such as being in a household with a malnourished person, no electricity, no access for clean water, shared sanitation .This result shows that the poverty status of rural households in Gozamen district is moderately poor. However in rural Ethiopia 96.30 percent of peoples are multidimensional poor, on average, the poor people are deprived 66.20 percents of the weighted indicators and the society is deprived in 63.7 percent of the total potential deprivation it could experience over all and hence, the multidimensional poverty status of rural Ethiopia is classified under extremely poor ((OPHI), 2017). The above table shows that the highest contribution to multidimensional poverty status of rural households in Gozamen District is standard of living deprivation which accounts 61.27 percent followed by education deprivation 26.24 percent and health deprivation 12.49 percent respectively. The data shows that standard of living take the largest domain of multidimensional poverty status of rural households in the sampled area of Gozamen district .The finding of the study confirm with . From the above table the highest share of multidimensional poverty highly comes from adult education, energy and floor having equal share, electricity, asset, clean water, sanitation, nutrition, child mortality and child education deprivation respectively. The above that show that in Gozamen district 82.16 percent of microfinance non-beneficiaries and 59.63 Microfinance loan service beneficiaries' peoples are multidimensional poor and the intensity of multidimensional poverty for microfinance loan service beneficial and non users of micro finance loan service were 46.32 percent and 50.4% respectively.

Decomposition of multidimensional poverty by access for microfinance service
The research result also shows that the non users of microfinance loan service beneficial were moderately poor i.e deprived in 41.41 percent and the member's society was deprived while users of microfinance loan service 27.62 percent deprivation of the total potential deprivation it could experience overall which shows that the users of the service were vulnerable to poverty or at risk of poverty. The chi2 test shows that there is an association between multidimensional poverty status and microfinance loan service at1% significance level. The above table shows that the highest contribution to multidimensional poverty index is standard of living for both microfinance loan service users and non users in Gozamen district which accounts about 68.91percent for microfinance loan service non users and 48.30 percent for microfinance loan service users.
The second highest contribution domain to poverty status of rural household in the study area is Education which accounts 30.20 percent for microfinance loan service non users and 19.38 percent for users.
The third contribution domain to poverty status of rural household is education which accounts 14.89 percent for microfinance loan service non users and 8.425 percent for microfinance loan service users. The data shows that standard of living is a serious problem for both in the sample of rural households. However, relatively the in all domains of multidimensional poverty the users of microfinance loan service is relatively lower than non-users.
Journal of Poverty, Investment andDevelopment www.iiste.org ISSN 2422-846X An International Peer-reviewed Journal Vol.53, 2020  The highest indicators of multidimensional poverty were adult education for both microfinance loan users and non users followed by floor. However, the lowest indicators of deprivation are child education for both microfinance loan service users and non -users. Despite these facts, the score of indicators of deprivation were relatively lower for users compare to that of non users.

Inferential statics
Before estimating the average treatment on the treated as a pre-requite the common support assumption were checked by kernel density plot, which ensures that there was a sufficient overlap in the characteristics of treated and non treated units to find adequate match which shows a lots of support between red and blue line (see appendix 2). Furthermore , the pstest were checked for balancing before trusting the ATT estimation and after matching, it was non-significant, so that the balancing was good for this study in building the good control group. The average absolute bias before matching was 8 and after matching it becomes 2.4 and hence the overall matching performance is good for all covariates ( see appendix 3).
Furthermore, Mantel Haenszel test statistics (MH) sensitivity analysis for average treatment effect were checked and there are no unobserved variable that affects treatment and the outcome variable simultaneously and hence, matching estimators are robust(see appendix ,4) . ATT .593103448 .770497457 -.177394009 .054530891 -3.25 Source: own survey, 2017. The above output shows that negative treatment effect on their multidimensional poverty head count of rural households(-0.177394009) difference is brought due to Microfinance loan service for users compare to that of non users and statistically significant at 1 percent level of significance. Alternatively, rural household of the treatment group, the treatment has reduced their multidimensional poverty head count by 0.177394009 on average.

The impact of microfinance loan service on poverty status of rural households in Gozamen District
The explanation is that microfinance loan service in Gozamen district allows them to purchase farm inputs, pity trade, animal fattening, current consumption, irrigation and horticulture activities which in turn leads to reduction on multidimensional status of household. This finding confirms to (Tsgay, 2014), (Adekola, G. and Dokubo,Chidinma, 2017) and (Eleuter Atilio Kihwele andRaphael Gwahula , 2015, Adams &Von Pischke, 1992;Coleman, 1999;Schicks & Rosenberg, 2011). As clearly shown in the table above the impact of microfinance loan service for standard of living is statistically significant that the individual of the treatment group, the treatment has reduced standard of living deprivation by -0.133177934 on average. The explanation is that the rural household who are users of microfinance loan service for the purchase of radio, television, and car and creates access for electricity, other assets and few energy sources which is a means for future production as well as raises their current consumption and their by reduce the standard of living dimension of poverty .

1) Standard of Living Dimension
2) Educational Dimension:-It consists of child education and adult education indicators of poverty. .011337335 -1.27 Source: own survey result, 2017. Note: ATT is average treatment effect on the treated The output shows that for the individual of the treated group, the treatment has reduced the educational deprivation by -0.014399756 on average. The explanation is that the availability of microfinance loan service allow the users to cover the costs of education and their by rise year of schooling and school attendance and their by reduce child education and adult education indicator of multidimensional poverty status of rural households.
3) Health Dimension:-It consists of two indicators of poverty i.e nutrition and child mortality. The output shows that for the households of the treated group, the treatment has reduced the health dimension of poverty by -0.04319011 on average. The explanation is that the prevailing loan service for users of the programme allows reducing child death and increasing expenditure on food, which in turn prevent infectious disease as well as improving nutritional status of children's and their by reduce the health dimension of poverty. The finding confirms (Nuredin Mohammed, Byeong Wan Le, 2015).For the validity of the average treatment effect.

SUMMARY, CONCLUSION AND RECOMMENDATION 4.1 Summary and Conclusion
This study found that 73.81 percent of rural peoples in Gozamen district are multidimensional poor, on average the poor people are deprived in 49.18 percents of the weighted indicators and the society is deprived in 36.30 percent of the total potential deprivation it could experience over all. Rural households in Gozamen district are deprived at least either all indicator of a single dimensions or a combination across dimensions such as being in a household with a malnourished person, no electricity, no access for clean water, shared sanitation however ,the poverty status of Gozamen district is classified as moderately poor.
The highest contribution to multidimensional poverty status of rural households in Gozamen District is standard of living deprivation which accounts 61.27 percent followed by education deprivation 26.24 percent and health deprivation 12.49 percent respectively. The data shows that standard of living take the largest domain of multidimensional poverty status of rural households in the sampled area of Gozamen district. The finding of the study confirm with Obadia,2014;Oluyombo,2013;Adekola,G.and Dokubo,Chidinma ,2017etc. The propensity score matching model result reveled negative treatment effect on their multidimensional poverty head count of rural households(-0.177394009) difference is brought due to Microfinance service intervention for users compare to that of non users and statistically significant at 1 percent level of significance.
It is also found that microfinance service has reduced standard of living deprivation by -0.133177934, health deprivation by-0.04319011 and education deprivation by 0.014399756 on average for treated compare to that of non treated group in Gozamen district. This finding confirms to (Odoyo, 2012), (Obadia, 2014), (Oluyombo, 2013) and (Adekola, G. and Dokubo,Chidinma , 2017)etc.

Recommendation
Based on this research finding, the researcher forwards the following recommendations; Microfinance loan service has a negatively impact on the multidimensional poverty status of rural household and hence, the government should give special attention to support microfinance those who support the rural poor household heads and improve the awareness level of farmers about it. Additionally, the microfinance financial institution expert should give attenstion for health, education and standard of living respectively to improve the dimension of deprivation multidimensional poverty. Furthermore, the rural households shall use microfinance loan service for health improvement, education improvements and standard of living improvements so as to reduce their multidimensional poverty in Gozamen district.