Daniel Haanwinckel is an Assistant Professor of Economics at UCLA. His research primarily focuses on the determinants of wages, unemployment, underemployment, and worker-firm sorting.

Abstract

How do different income taxation systems---for instance, individual vs. joint taxation---affect household decisions and welfare? We provide an answer to this question in three steps. First, we document significant labor supply disincentives for secondary earners in the U.S. taxation system for married couples and show that secondary earners respond to these disincentives. Next, we develop a lifecycle model with labor supply choices on the intensive and extensive margins, endogenous marriage and divorce decisions with limited commitment, assortative mating on both education and unobserved ability, production of a household consumption good using time and market inputs, and human capital accumulation on the job. We estimate the model using PSID, CPS, and ATUS data and validate its predictions on the labor supply effects of income taxation using the 2003 ``Bush Tax Cuts.'' Finally, we use the model to compare four tax systems: an ``income-splitting'' system similar to the one adopted by the U.S., an individual taxation system, a more general joint taxation system that allows for arbitrary ``jointness,'' and an income-splitting system with a flat deduction on the secondary earners' income. We confirm previous findings that the optimal individual system welfare dominates the optimal income-splitting system. However, we also document that the optimal general joint system generates significantly higher welfare than both. The main reason is that the general system allows for greater progressivity, as the individual one, but still manages to generate significant ``marriage tax bonuses'' that are particularly beneficial to low-education, low-ability individuals. Lastly, we document that the income-splitting system with a secondary-earner deduction generates more welfare than the individual system while maintaining simplicity and parsimony.

Authors: Mary Ann Bronson (Georgetown University), Daniel Haanwinckel (UCLA and NBER), Maurizio Mazzocco (UCLA)