Using a bivariate probit model to control selective access to employment, this paper tests the hypothesis that the magnitude of the negative impact of geographical externalities on the risk of overeducation is higher for male international immigrants, as they look for jobs in the global labour market, and lower for females, as they are  ‘tied stayers’ or  ‘tied movers’ . The results show that (a) the per-capita wage of a region where an employed person lives, its educational endowment and the average per-capita wage of its neighbouring regions can explain the phenomenon of overeducation, and (b) these effects are moderated by the combined influence of migration status (locals, intranational immigrants and international immigrants) and gender. However, the sign and the scale of the coefficients on externalities seem to reject the hypothesis. This seems to suggest that male migration in regions that are full of externalities is not a response to overeducation and women job-seeking is not restricted to the market chosen by her partner.
Key words: overeducation, externalities, migration, gender