Although prior studies of female labor supply in developing countries use individual
cross-sectional data in the estimation, such data cannot capture the behavior of labor over time
(Contreras, et al. 2005), and estimates may be biased due to omitted individual effects. Panel
data allows for control of unobserved individual effects, but such data sets are seldom available
in developing countries. In the absence of panel data, Deaton (1985) proposes the construction of
synthetic cohorts, or pseudo-panel data, from the time series of cross-sectional surveys. By
averaging over individuals within each cohort to eliminate individual effects, this approach can