Migration distances vary widely across individuals and contexts, yet empirical evidence on the drivers of migration distance within Africa remains limited. While much of the migration literature focuses on Western destinations, less is known about how far migrants travel within the African continent and what factors are associated with these trajectories. This study addresses this gap using a large micro-level dataset collected by the International Organization for Migration between 2021 and 2023, comprising over 60,000 land-based migration observations from eight West African countries. We pursue two objectives. First, we assess the extent to which intended migration distance can be predicted using information available at transit points. Second, we identify and rank the key individual and contextual predictors associated with intended migration distance using interpretable machine learning and statistical models. The analysis reveals a bimodal distribution of migration distances: most individuals move locally, typically within 100 km, while a smaller yet significant share undertake long-distance journeys exceeding 1500 km. Employment status, GDP (at the residence locality) and reasons for travel are among the most influential predictors of intended migration distance. Unemployed migrants tend to travel significantly farther than employed individuals, suggesting persistent economic constraints and opportunity-seeking behavior. In contrast, conflict-related mobility shows marked temporal variability with a sharp increase in long-distance migration in 2023, coinciding with intensified violence in key origin countries. These patterns suggest an asymmetry between economic and conflict-driven mobility: economic pressures are associated with relatively stable distances traveled, whereas conflict introduces greater volatility and route reconfiguration. By combining rare field data with interpretable machine learning, this study advances quantitative understanding of mobility dynamics in West Africa and provides evidence relevant to migration forecasting and humanitarian policy in underrepresented regions.
Tafani et al. (Fri,) studied this question.