The effect of a child on female work when family planning may fail

This paper develops a structural empirical model of contraception and participation choice under imperfect control of fertility and learning in order to estimate the effect of the first born child on female work. I exploit family planning failures as sources of variation using the 1995 US National Survey of Family Growth (NSFG) wich contains full retrospective information on participation, contraception and children. Key factors driving the importance of the effect are education, labor market experience, child´s age and preferences for leisure and children. Additionally, I estimate Average Treatments Effects (ATE) of the first (-12.4%), second (-5.6%) and third born children (-4.9%) using as additional sources of variation twins and the sex of the first two children. Finally, I propose a weighting procedure to estimate Local Average Treatment Effects (LATE).