We estimate the causal effect of additional childbirth on mothers’ probability of working by using the two-stage least squares estimation method. Samples of mothers drawn from the 10%–20% Restricted Access Samples of the 2000, 2005, 2010, and 2015 Censuses are used. We use children’s gender composition as the instrumental variable for childbirth. We find that additional childbirth had significantly adverse effects on mothers’ probability to work up to 2010, but it is no longer likely to be the case. The ordinary least squares estimates still show a strong negative correlation between childbirth and maternal labor supply, which, our results indicate, is driven mainly by endogenous self-selection. We find that the labor supply of less-educated mothers is more adversely affected by childbirths.