Analysis of pseudo-panel data with dependent samples

In this paper, we discuss a model for pseudo-panel data when some but not all of the individuals stay in the sample for more than one period. We use data on the labor market of the Basque Country from 1993 to 1999 treated through FORTRAN 77 programing. We construct economically reasonable age cohorts for active population and use gender, qualification and social status as explanatory variables in our model. Given the class of data we use, we analyze the properties of the random error and estimate the model through maximum likelihood, finding significant results from an applied point of view.