Modeling the Impact of some Macroeconomic Variables on Unemployment Rate in Africa

Abstract

Unemployment remains a challenging socio-economic problem in Africa. This study used panel data to investigate the impact of some macroeconomic variables on unemployment rate in Africa with data from World Development Indicators ranging from 1994 to 2023 on 49 Africa countries. Traditional panel models and machine learning models were implemented. The Hausman test revealed that, the fixed effect model was the best traditional model and the multilayer perceptron model was recorded as the overall best model. From the MLP model, the three most significant variables that affect the unemployment rate are agriculture, population growth, which have a negative relationship, and current account balance, with a positive relationship. The study recommends that better policies, such as startup capital for youth in agriculture, and modern machines should be made available, incentives for a few families, and restrictions on the maximum number of children a family can have to regulate population growth.