Incorporating living conditions and job opportunities in cities into mathematical models of human mobility improves model accuracy. The traditional gravity model of human mobility uses the distance of a move and the population of a destination city to predict migration patterns, with larger cities exerting more “pull” than smaller cities. The competing radiation model is based on quantifying the opportunity available in a destination location. Maurizio Porfiri and colleagues added to the radiation model, weaving in measures for living conditions and job quality. These measures include the presence of conflicts, natural hazards, and political persecution, and income or wealth inequality. Differences in these factors can create inequalities between locations that can repel or attract. The authors present results on migrations in South Sudan, which are driven by conflicts when considered in combination with natural hazards such as floods. The authors also present results on commuting patterns in the United States, where movements are better predicted when one includes as variables the fraction of household income spent on rent, the proportion of people living in severe poverty, and income inequality. Although there’s still a gap between predictions and reality, in both cases, the authors’ proposed model outperforms a standard radiation model. According to the authors, a model sensitive to inequalities between the quality of locations will improve migration forecasting, especially as climate-related migrations increase.
PNAS Nexus
Predicting the role of inequalities on human mobility patterns
20-Jan-2026