ON THE ESTIMATION OF A FIXED EFFECTS MODEL WITH SELECTIVITY BIAS

In case of sample selectivity the maximum likelihood estimator of the parameters in a model with fixed effects will not be consistent when the number of time periods is small. In this paper, we present a transformation to eliminate the fixed individual effects and show that the corresponding marginal maximum likelihood estimator is computationally feasible and can be used to estimate the remaining parameters consistently even if number of time periods is finite. © 1990.