Child penalty estimation and mothers' age at first birth

We show that the widespread approach to estimate the career costs of motherhood - so called "child penalties" - is prone to produce biased results, as it pools first-time mothers of all ages without accounting for their differences in characteristics and outcomes. We propose a novel method building on the recent advances in the difference-in-differences literature to address this issue. Applied to German administrative data, our method yields 30 percent larger post-birth earnings losses than the conventional approach. We document meaningful effect heterogeneity by maternal age in both magnitude and interpretation, highlighting its key role in understanding the impact of motherhood.