Estimating Labor Supply Response

Recent cross-sectional labor supply studies have yielded a wide range of parameter estimates; they have therefore been of little use to policymakers attempting to evaluate welfare reform proposals. In this report a step-by-step empirical exploration of alternative labor supply estimating equations is undertaken to identify the independent (marginal) effects of changes in sample composition, variable definition, equation specification, and estimation technique. Samples used consist of white, married, male family heads, aged 25-54, from the Survey of Economic Opportunity. The findings indicated that major differences in wage and income effects estimated from cross-sectional data can often be attributed to (1) the measure of labor supply used, (2) the broadness of the sample used, especially the inclusion of individuals with zero values for the dependent variable, and (3) for those studies that used imputed wages, the exclusion or inclusion of education in the labor supply regression.