TY - JOUR AU - Kleven,Henrik Jacobsen AU - Kreiner,Claus Thustrup AU - Saez,Emmanuel TI - The Optimal Income Taxation of Couples JF - National Bureau of Economic Research Working Paper Series VL - No. 12685 PY - 2006 Y2 - November 2006 UR - http://www.nber.org/papers/w12685 L1 - http://www.nber.org/papers/w12685.pdf N1 - Author contact info: Henrik Kleven Department of Economics & STICERD LSE Houghton Street London WC2A 2AE United Kingdom E-Mail: H.J.kleven@lse.ac.uk Claus Kreiner Institute of Economics University of Copenhagen Studiestraede 6 DK-1455 Copenhagen K Denmark Tel: 4535323020 Fax: 4535323000 E-Mail: claus.thustrup.kreiner@econ.ku.dk Emmanuel Saez Department of Economics University of California, Berkeley 530 Evans Hall #3880 Berkeley, CA 94720 Tel: 510/642-4631 Fax: 510/642-6615 E-Mail: saez@econ.berkeley.edu AB - This paper analyzes the optimal income tax treatment of couples. Each couple is modelled as a single rational economic agent supplying labor along two dimensions: primary and secondary earnings. We consider fully general joint income tax systems. Separate taxation is never optimal if social welfare depends on total couple incomes. In a model where secondary earners make only a binary work decision (work or not work), we demonstrate that the marginal tax rate of the primary earner is lower when the spouse works. As a result, the tax distortion on the secondary earner decreases with the earnings of the primary earner and actually vanishes to zero asymptotically. Such negative jointness is optimal because redistribution from two-earner toward one-earner couples is more valuable when primary earner income is lower. We also consider a model where both spouses display intensive labor supply responses. In that context, we show that, starting from the optimal separable tax schedules, introducing some negative jointness is always desirable. Numerical simulations suggest that, in that model, it is also optimal for the marginal tax rate on one earner to decrease with the earnings of his/her spouse. We argue that many actual redistribution systems, featuring family-based transfers combined with individually-based taxes, generate schedules with negative jointness. ER -