Events at CERGE-EI

Michele Aquaro

University of Cambridge, United Kingdom

Authors: Michele Aquaro and Pavel Čížek

Abstract: A new estimation procedure of dynamic panel data models with fixed effects is proposed. To improve upon existing estimators, we propose to apply the pairwise-difference data transformation to the generalized method of moments based estimators. A particular focus is given to the long difference (LD) estimation procedure of Hahn et al. (2007), which was proved to retain strong moment conditions even when data are persistent without imposing further assumptions. The bias and asymptotic distribution of the original LD estimator and its proposed extensions are derived. A simulation study is conducted to assess the finite-sample properties of the estimators.

JEL code: C13, C23, C26

Keywords: asymptotic distribution, dynamic panel data, generalized method of moments, long difference, pairwise differences

Full Text:  “Pairwise difference estimation of dynamic panel data models”