Travail de recherche/Working paper
Résumé : Labor automation is generally associated with a decrease in demand for mid-skill jobs,often routine-intensive, in favor of the others. This paper investigates its effects onfertility timing decisions using European panel data, by constructing a measure of localexposure to industrial robotics, and by adopting a Fixed Effect with Two-StageLeast Squares methodology. Higher exposure is associated with an anticipation offertility in low- and high-skilled regional labor markets, and with its postponementin medium-skilled ones. An optimal stopping model, in which individuals adjust thetiming based on their future labor opportunities, formalizes the causal intuition. Itsnumerical application, based on survey data, suggests that the effect of an increase inobserved automation on the willingness to postpone fertility is concave with respect toeducation, consistently with the Routine-Biased Technological Change hypothesis.