Lifecycle Wages and Human Capital Investments: Selection and Missing Data

Working Paper: CEPR ID: DP16999

Authors: Laurent Gobillon; Thierry Magnac; Sébastien Roux

Abstract: We derive wage equations with individual specific coefficients from a structural model of human capital investments over the life-cycle. This model allows for interruptions in labor market participation, and addresses missing data and attrition issues. We further control for selection in a flexible way by using interactive effects. Estimation is based on long administrative panel data of male wages in the private sector in France. A structural function approach shows that interruptions negatively affect average wages. More surprisingly, they also negatively affect the inter-decile range of wages after twenty years, and this is due to interruptions being endogeneous. These results question the popular Missing At Random assumption that is made when assessing the building up of wage inequalities over the life cycle.

Keywords: human capital investment; wage inequalities; factor models; missing data

JEL Codes: C38; D91; I24; J24; J31


Causal Claims Network Graph

Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.


Causal Claims

CauseEffect
interruptions in labor market participation (J22)average wages (J31)
interruptions in labor market participation (J22)interdecile range of wages (J31)
interruptions in labor market participation (J22)wage inequality (J31)

Back to index