MODELING POVERTY STATUS OF GHANAIANS: A LOGISTIC REGRESSION APPROACH

ABSTRACT The poor are not evenly distributed within the country and they do not share the same socio-economic and demographic characteristics. It is against this  background that analysis of the characteristics  that differentiate the poor from  the  non-poor  in  Ghana  cannot  be  underestimated.  Poverty  indicators make it possible to analyze the likely  determinants and are, therefore, essential for formulating  policy  interventions that may contribute  directly or indirectly to its alleviation. This study therefore aims at determining the most important characteristics that differentiate poor households from non-poor households  in Ghana.To achieve this objective, data was obtained from the Ghana Statistical Service.  It  consists  of  49,005  households  surveyed  in  the  country  of  which 17.2 percent were classified  as  being poor (extremely poor). The analysis of the  data  relied  mainly  on  logistic  regression.  It  was  found  that  the  region  a household  resides  in,  the  number  of  persons  in  that  household,  access  to improved  sources  of  water  and  sanitation,  are  the  most  important characteristics. These and other interesting results are discussed.