Analysing the impact of individual factors on local employment-education matching across Tunisian regions

The aim of the study is to investigate spatial relationship between individual and socio-economic determinants and education-employment match in Tunisia with unique data on 268 delegations and 162 sectors of Tunis district in 2024. The article demonstrates that spatially explicit regression techniques provide a deep understanding of the spatial heterogeneity underlying education–employment mismatches in Tunisia compared to conventional non-spatial approaches. A Geographically Weighted Logistic Regression (GWLR) model and its multi-scale extension (MGWLR) were estimated to account simultaneously for spatial non-stationarity and varying local relationships between explanatory variables—such as participation in vocational insertion programs, university type, and marital status. Model comparison results indicate that both GWLR and MGWLR outperform the global logistic regression model in terms of local fit and predictive accuracy, revealing strong spatial disparities in the returns to education and training. These findings highlight the importance of spatially differentiated employment policies and region-specific human capital strategies to reduce labor market mismatches in urban Tunisia. JEL classification : J41, C31