Proceedings of the 52nd International Academic Conference, Barcelona

ROBUST STATISTICAL INFERENCE IN PANEL DATA INCLUDING A PRACTICAL APPLICATION IN THE MIGRATION ANALYSIS

TATIANA MEDKOVÁ

Abstract:

The need for robust statistical inference is well-documented even in the elementary case of a regression with a randomly sampled cross section. The usual ordinary least square standard errors are generally biased under the presence of heteroskedasticity; a phenomenon that seems to be a rule rather than an exception in applied anaylsis. The article describes several methods to deal with the biased standard errors grouping them in two categories: sandwich variance estimators and multi-way clustering. Moreover, the empirical application is included. An analysis of migration in European countries using the theory of gravity model is done applying several standard errors correction methods.

Keywords: migration, robust, standard errors, heteroskedasticity, panel data

DOI: 10.20472/IAC.2019.052.039

PDF: Download