## Implications for Poverty Reduction in Rural Households in Ghana: Results

The results and discussion on the estimated treatment effect model is presented in this section. This is preceded by a presentation of the descriptive statistics of the variables used for the empirical estimation. Table 1 shows the description of the variables used for the analysis. Age, depend and hhppl are a set of household characteristics denoted by rin the model. Where age is number of years of the respondent, the mean age is given as 39.95 or approximately 40 years. The number of dependents in the respondent’s household is denoted by depend with a mean number of three dependents. Also hhppl denote the number of people in the respondent’s household with a mean of about 7 people in each household. The variable frnsours denotes the number of friends of the respondent who have borrowed from an MFI. Direct marketing

Also, lonsours measures the number of sources of borrowing that the respondent can actually borrow from within the community when in need of a loan. These sources include both formal (MFIs) and informal (friends, relatives and money lenders) institutions. Thus lonsours determines the number of these formal and informal financial institutions that the respondent can actually and confidently go to for a loan when in need. wkexpend2 is the weekly consumption expenditure of the respondent on basic needs (food, clothes and rent) and this is denoted by C in equation. The cost of the basic needs was valued at prices in Bolgatanga.
The amount of profit made in a month is denoted by amtprof (x). The current value of assets owned by the respondents measured in Cedis is denoted by assets. The variable oldsav ( R1 ) measures the initial financial resources or savings of the respondent before receipt of loan from a MFI or start of agro-processing business. Again, acesmf (access to MF), is a dummy variable which takes the value 1 if the respondent has received a loan from a MFI. The mean value of acesmf is 0.508; this means that 51% of the respondents received at least a loan from an MFI. The variable numacty denotes the number of income generating activities that the respondent engages in as at the time of the study.
Kasena, bwest, builsa, talensi and bongo are district dummy variables for Kasena/Nankani, Bawku West, Builsa, Talensi/Nabdan and Bongo Districts respectively. The mean value for each of the Districts is given as 0.202, 0.198, 0.198, 0.202 and 0.202 respectively for Kasena/Nankani, Bawku West, Builsa, Talensi/Nabdan and Bongo Districts. This suggesting about 20% of the respondents are from each of the five Districts. However, Kasena/Nankani and Bawku West Districts have very vibrant market centres with high patronage from neighbouring country, i.e., Burkina Faso. Therefore, the two spatial dummy variables were employed in the empirical model to capture the two major market areas. The remaining Districts; Builsa, Talensi/Nabdan and Bongo Districts which do not have vibrant market centres were used as the reference categories.
Also, the variable edulev is a categorical variable; it measures the highest educational level of the respondent. The mean of each level (category) of education shows the proportion (percentage) of the respondents in that category. Thus out of this variable, 53.1% have no education; 24.5% have Primary school education; 15.1% have JSS/Middle school education; 6.6% have Senior Secondary School (SSS)/ secondary school education while 0.7% have post secondary school education. Respondents with no education are used as the reference category.

Table 1: Descriptive Statistics of Variables used for the study

 Variables Description Obs. Mean S.Deviation kasena Kasena Nankana District (1/0) 437 0.202 0.402 bwest Bawku west District (1/0) 437 0.198 0.398 builsa Builsa District (1/0) 437 0.198 0.40 Talensi Talensi/Nabdan District (1/0) 437 0.202 0.402 bongo Bongo District(1/0) 437 0.202 0.402 age Age in years 437 39.951 11.659 depend Number of dependants in household 437 3.100 2.121 frnsours Number of friends with loans 437 2.995 4.547 hhppl Number of people in household 437 7.032 7.032 acesmf Has received MFI loan(1/0) 437 0.508 0.500 wkexpend2 Weekly consumption expenditure 437 2.984 0.603 lonsours Number of borrowing sources 437 1.062 0.381 amtprof Amount of profit 437 41.732 52.579 numacty Number of economic activities 437 1.245 1.123 oldsav Initial savings 457 98.993 133.701 assets Value of physical assets 437 442.716 1040.845 edulev Educational level: 437 No education (1/0) 0.531 0.500 Primary school(1/0) 0.245 0.430 JSS/Middle school(1/0) 0.151 0.358 SSS/Secondary school(1/0) 0.066 0.249 Post Secondary school(1/0) 0.007 0.0827