POVERTY DYNAMICS IN KAGERA REGION a- 1991-2010

This study examines the dynamics of poverty in Kagera Region during 1991-2010 period using the Kagera Health and Demographic Survey panel data. First, the study decomposes the Kagera headcount poverty index into growth, inequality and rural/urban population shift effects using Son (2003) approach. The results show population shifts was pro-poor and growth led to reduction of poverty within the region. These results imply that there is need for pro-poor employment creation strategies in urban areas and increased productivity in rural areas. Second, employing the fixed effects model and dynamic panel data estimators and borrowing from Ravallion (1988) approach, the study examines the dynamics of householda€˜s consumption and vulnerability. The dynamic panel data estimators show that Life Cycle Hypothesis for consumption smoothing does not hold, whereas the Ravallion (1988) approach show that risk to poverty is less for the 1991 poverty line and high for 2010 poverty line. Thus, pro-poor farming approaches and earmarking resources for consumption stabilization interventions are needed. Lastly, the study examines a householda€˜s duration in poverty and factors accounting for falling into or moving out of poverty. The non- parametric and semi-parametric discrete data spell approaches are employed. The non-parametric results show that poverty is transitory, and that male-headed households have a higher likelihood of exiting poverty than female-headed ones. The duration variables have significant positive effect for exiting poverty. In this regard, policy may entail enhancing social networks and improving the balance in resource distribution and allocation between sexes.