Unlike the standard cross-country growth regressions that focus on the long run average growth and hence mask episodes of high and low growth that most of these countries experience, this paper follows the recent literature and examine growth episodes in SSA countries. This new approach applied enable us identify years of growth accelerations and episodes which were sustained over the medium and long-term. More specifically, utilizing the Penn World Data Tables (version 6.2), the study follows the diagnostic or two-pronged strategy of Rodrik (2005a and 2005b), which focuses on the particular constraints that prevent a given country from growing faster. In the latter part, we analyze the impact of growth on income poverty and inequality in selected countries drawing on available household survey data published by the World Bank in its Global Poverty Monitoring Database.To establish correlation between growth episodes(all accelerations and sustained ones) and policy variables, institutional variables and geographic factors, we estimate alternative limited dependant variable models. The results show that variables affect these two growth episodes differently.