Governance and poverty in sub-saharan Africa: a random forest approach

Abstract Our study applies Random Forest classifier which challenges Logit to investigate the relationship between governance and multidimensional poverty in Sub-Saharan Africa using Afrobarometer 2022 data from multiple countries. The findings reveal that negative perceptions of governance such as high perceived corruption, dissatisfaction with public services, disapproval of political leadership significantly increase the probability of falling into higher poverty categories, while factors like higher education, urban residence, and access to public services reduce it. The results suggest that improving governance transparency, investing in education and health services, and promoting inclusive policies targeting rural and vulnerable populations are critical to reducing poverty. Classification JEL : I30, O15, C60