Panel data modeling is a problem that arises across several areas, namely demography, economics, finance, biology, climatology and environment. In empirical studies, the estimation of the parameters of these models is usually done based on classical estimation methods. The presence of outliers in panel data is a frequent situation in data sets. real data in different areas and can drastically affect estimates obtained by classical estimation methods. The objective of this work is to develop estimation techniques for models with panel data that are robust, that is, that generate better estimates than classical estimators, when the data do not verify the assumptions necessary to validate the properties of classical estimators. The robustness properties of the proposed procedures are investigated through simulation studies, under different scenarios. To illustrate the performance of the proposed robust methods, an example is also presented, based on a set of real economic panel data, known from the literature.

published