Working Paper: NBER ID: w1638
Authors: William T. Dickens; Shelly J. Lundberg
Abstract: This study presents a model of labor supply in which individuals may face constraints on their choice of work hours, and analyzes the sensitivity of parameter estimates and policy conclusions to the usual assumption of unrestricted choice. We set up the labor supply decision asa discrete choice problem, where each worker faces a finite number of employment opportunities, each offering fixed hours of work.The distribution from which these are drawn, as well as the number of draws, is estimated along with the behavioral parameters of individual labor supply.The standard model with unconstrained hours appears as a special case where the number of draws approaches infinity. We estimate the mean absolute difference between desired and actual work hours to be about ten hours perweek. The results strongly support the notion that hours choices are constrained, and suggest that models which ignore restrictions on hours worked may yield biased estimates of the wage elasticity of desired hours. Further, we suggest that analysis of policies such as income transfers and the flat rate tax which do not consider their effects on the distribution of hours offered may be very misleading.
Keywords: Labor Supply; Hours Restrictions; Discrete Choice Model
JEL Codes: J22; D12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Job offer distributions (D39) | Actual hours worked (J22) |
| Unrestricted choice in hours worked (J22) | Biased estimates of wage elasticity of desired hours (J38) |
| Ignoring constraints on hours (C41) | Misleading conclusions about tax and transfer systems on labor supply (H31) |