Analysis of panel data models with grouped observations

Abstract

We present an iterative estimation procedure to estimate panel data models when some observations are missed or grouped with arbitrary classification intervals. The analysis is carried out from the perspective of panel data models, in which the error terms may follow an arbitrary distribution. We propose an easy-to-implement algorithm to estimate all of the model parameters and the asymptotic stochastic properties of the resulting estimate are investigated as the number of individuals and the number of time periods increase.