Polarized education levels and civil unrest

After introducing a measure for educational polarization (EduPol ), this paper presents a theoretical framework to understand whether and how EduPol may affect the contest for power in society. The model suggests that societies with high degrees of EduPol (i.e., substantial shares with either no or university-level education) are systematically more prone to civil unrest. We test this prediction on four measures of civil unrest: Political instability, domestic terrorism, civil conflict, and civil war. Our empirical estimations produce evidence consistent with this hypothesis as all four phenomena are positively associated with EduPol at the beginning of the respective period, exhibiting meaningful magnitudes. These results prevail when accounting for (i) potentially confounding factors, (ii) country- and time-fixed effects, (iii) economic inequality, (iv) ethnic and religious polarization and fractionalization, and (v) numerous alternative estimations and outcome variables.