Optimal Taxation of Families : Mirrlees Meets Becker

This paper studies the optimal taxation of families in an environment in which (i) potential parents differ by their labor productivity and child-bearing preference, which are only observable by the family, and (ii) child-rearing requires both goods and parental time. Potential parents simultaneously decide how much labor to exert and how many children to have. The government uses information on family income and size to construct an optimal tax system through the combination of an income tax schedule and child tax credits. I observe that parental time and the cost of goods involved in child-rearing have distinct impacts on the shape of optimal child tax credits. As a quantitative analysis, I calibrate my model to US data and show that childrearing costs translate into a pattern of optimal credits that is U-shaped in income. In particular, tax credits decrease along the first nine deciles of the income distribution and increase thereafter. In addition, tax credits decrease according to family size because of the economies of scale in child-rearing. For median-income families, the credit for the second child equals 71% of the credit for the first child. I also find that implementing the optimum yields large welfare gains