Yuanyuan Chen *
May 8, 2005
Abstract:This paper examines the extent to which non-random selection in labor market affects assessing the black-white wage difference by applying a longitudinal method of imputing wages for nonworkers. Since fixed effects are crucial in determining the wages of nonworkers, the cross-sectional methods of controlling only on observables understate the impact of selection. Using the Panel Study of Income Dynamic (PSID) 1970 - 2000 data, I find that the selective bias can explain 40 percent of the observed change in the black-white wage gap over 1970-2000.
pdf Download
* Author's Email Address: ychen@imf.org
I am grateful to Peter Gottschalk, Author Lewbel, Donald Cox, Shang-jin Wei and all participants in the seminar at Boston College Economics Department for helpful guidance and comments. All errors are mine.