Duke Economics Working Paper #03-22
Several approaches have been proposed to estimate family models of labor supply in the presence of taxes. They all, however, either require closed functional forms for both unconditional and conditional labor supply functions, or require the discretization of the budget constraint. This, in turn, either limits the choices for functional forms, necessitates undesirable assumptions about the behavioral model or the stochastic specification, or both. I propose an estimation method that suffers from none of these restrictions. By estimating parameters of a joint utility function and employing numerical methods, this estimation method allows for continuous hours choices of the husband and wife, and incorporates measurement error in hours and heterogeneity in tastes for work. This specification also facilitates straightforward incorporation of fixed costs of work. Using data from the 2001 wave of the Panel Study of Income Dynamics, I estimate parameters of a joint utility function over huband's and wife's hours of work and total consumption. I then use these results to simulate the effects of the Tax Relief Acts of 2001 and 2003 on married couples' labor supply, tax revenue, and the deadweight loss of the income tax. These simulations predict large negative tax revenue effects of TRA 2001 and TRA 2003, but very small effects on the labor supply of married couples and on the deadweight loss of the income tax.
JEL:
Retrieve document:
43 pages