Maternity leave policies are presumed to be essential to ensure the health of pregnant workers and their unborn children. However, little is known about the optimal duration of prenatal maternity leave and existing policies are not evidence-based. We evaluate a substantial maternity leave extension in Austria, which increased mandatory leave prior to birth from six to eight weeks. Our estimation strategy exploits that the eligibility for the extended leave was determined by a cutoff due date. As an additional source of exogenous variation, we use information on non-working mothers, who are not eligible for maternity leave. Across estimations, we consistently find no evidence for significant effects of this extension on children's health at birth, subsequent maternal health and fertility, and longterm human capital outcomes of children. Our finding is confirmed by a supplementary cross-country panel analysis.