The study uses administrative data from Luxembourg to investigate fathers' decisions to use parental leave. We focus on two measures of opportunity cost: the difference between the parental leave benefit and the salary of the father and the mean salary growth for a period of 6 months for each father. The first measure captures the direct opportunity cost, while the second is a proxy for foregone promotion opportunities. We use Cox proportional hazards model for the analysis. The results suggest a negative relationship between foregone income and taking parental leave. Surprisingly, salary growth appears to be positively related to the hazard of taking parental leave. Coefficients of control variables are in line with previous findings: fathers are more likely to use parental leave if they work in larger organization and for the first child.