Acceso usuarios
Ignacio Moral Arce, Concepció Patxot Cardoner, Guadalupe Souto Nieves
In this paper we estimate lifetime earnings profiles for Spain using a new dataset based on administrative Spanish Social Security records from 1980 to 2005. In order to exploit the panel data dimension of this dataset, we use panel data estimates specially derived for censored data as is the case. First, the Kalwij (2003) estimator derived for censored panel data is used in order to estimate the censored earnings, controlling for some identifiable variables. Second, to overcome the main weakness of the above-mentioned method, i.e. the impossibility of identifying variables that remain constant over time, the Wooldridge (1995) and Zabel (1992) estimators are used. Finally, given the difficulties in isolating the period, age and cohort effects, linked by a linear relationship, a test of uniform wage growth is performed following Fitzenberger et al. (2001). As the results of the test indicate that the age and period (or cohort) effects can be considered as separable, a tentative age profile can be derived for each gender and contribution group.
