Can moving to an earnings-related parental leave system influence children's wellbeing and are heterogeneous effects on parents carried over to the entire family, making special groups of children worse off than others? To answer this question, this study exploits a large and unanticipated parental leave reform in Germany as a natural experiment. By replacing a means-tested by an earnings-related system the reform affected different groups of families to a variable extent. I detect significant negative effects on the personality of newborns whose families are subject to a nonpositive change in the overall benefit amount compared to the pre-reform situation. 2-3-year-old children belonging to the reform's winners, however, improve their basic life skills and language skills.