## Implications for Poverty Reduction in Rural Households in Ghana: The Level of Education

Similarly, the number of sources of borrowing that the respondent can actually borrow from within the community when in need of a loan (lonsours) has a negative association with the probability of taking a loan from an MFI. Given the coefficient of lonsours as -0.331, then an increase in the number of sources one can actually borrow from, both formal and informal MFIs, reduces the probability of taking a loan from an MFI. A possible reason is that, proliferation of sources of borrowing both formal and informal in rural communities has the tendency of exploiting borrowers especially with their (lenders) inclination to sustainability by charging high interest rates so as to meet their operating cost. This point has been well reiterated in Annim. This assertion was confirmed by majority of the non-beneficiaries as they cited among other reasons, high interest rates charged by the MFIs as a reason why they have not taken a loan from an MFI. This brings to the fore the question of sustainability of MFIs particularly those operating in the rural areas as against the goal of poverty reduction. CRM analytics

The amount of profit respondents make in a month (amtprof) is positively associated with the probability of borrowing from an MFI. Thus as amtprof increases given the coefficient of 0.00396, the probability of taking a loan also increases. As one makes more profits from her business she is motivated thus increasing the probability of her going in for another loan. The number of friends of the respondent who have ever borrowed from an MFI (frnsours) is also positively associated with the probability of borrowing from an MFI. frnsours measures the dept of MFIs serves, thus as more rural poor people are served by the MFIs, the probability of those served encouraging their friends or peers to also borrow from the MFIs increases given its coefficient of 0.298. From Table 3b, the log of Weekly consumption expenditure (lwkexpend2) is explained by Kasena Nankani District (kasena), Bawku West District (bwest), age (age), age-squared (age1), Primary school category of the level of education and access to MF (acesmf) as all are statistically different from zero at 1% significance level. Also lwkexpend2 is explained by depend, amtprof, JSS/Middle and numacty, these are statistically different from zero at 5% significance level. Again, oldsav is significance at 10%.

This suggests that with kasena, given its coefficient as 0.342 holding all other variables constant, respondents from the Kasena Nankana District spend 34% per week on basic needs more than respondents from Talensi Nabdan, Bongo and Builsa Districts which are the reference categories. The coefficient of bwest is 0.548, implying that respondents from the Bawku West District also spend on the average 55% per week on basic needs more than those

from Talensi Nabdan, Bongo and Builsa Districts which are the reference categories. However respondents from the Bawku West District spend about 21% per week more than their counterparts in the Kasena Nankana District. Thus the Bawku West and the Kasena Nankana District contribute to increased consumption expenditure for the respondents in these two Districts as opposed to the other three Districts contribute to respondents from those Districts.

Both the Kasena Nankana and Bawku West districts have very vibrant market centres with patronage from neighbouring country Burkina Faso. It is possible that products of the agro-processors in these two Districts enjoy good demand from their respective market leading to high profits which is ultimately translated into high consumption levels. This finding therefore concurs with Kiiru & Mburu, where it was argued that microfinance can not improve welfare unless there is effective demand for goods and services, which ensures that the products of micro-entrepreneurs are consumed. These Districts have also had numerous MFIs particularly the Bawku West District. This could explain the reason why that District has such high percentage consumption expenditure per week (55%).