Teachers' sick leave and summer timing: an impact evaluation

In the context of the growing share of public workers in the Spanish labour force, this paper analyses opportunistic behaviour in sick leave duration and return-to-work probabilities among teachers. We use administrative microdata (2011-2019), transformed into an original monthly spell-level panel. The analysis employs a tailored regression within a difference-in-differences framework exploiting the academic calendar and comparing professional sick leaves among pre-university teachers in public and private schools. Results show that the timing of sick leave endings systematically responds to the school calendar. Public-school teachers display longer sick leave durations and lower return-to-work probabilities than private-school teachers, consistent with ex post moral hazard under more generous employment protection and sick leave coverage. Around the summer break, these differences become more pronounced: public-school teachers extend their leaves further relative to private-school teachers. Simultaneously, the usual gap in return-to-work probabilities temporarily reverses, with public-school teachers becoming more likely to end their leave immediately after summer. These patterns are especially strong among mid-career workers, women, and cases involving more subjective or difficult-to-diagnose injuries.