A0177
Title: A projection based approach for interactive fixed effects panel data models
Authors: Juan Manuel Rodriguez-Poo - Universidad de Cantabria (Spain) [presenting]
Georg Keilbar - Humboldt-University of Berlin (Germany)
Alexandra Soberon - Universidad de Cantabria (Spain)
Weining Wang - University of York (United Kingdom)
Abstract: The aim is to present a new approach to estimation and inference in panel data models with interactive fixed effects, where the unobserved factor loadings can be correlated with the regressors. A distinctive feature of the proposed approach is to assume a nonparametric specification for the factor loadings, which allows us to partially out the interactive effects using sieve basis functions to estimate the slope parameters directly. The new estimator adopts the well-known partial least squares form, and its consistency and asymptotic normality are shown. It is found that the limiting distribution of the estimator has a discontinuity when the variance of the tcharacteristictic parameter is near the boundaries, which makes the usual "plug-in" method used to estimate the asymptotic variance only valid pointwise and may produce either over- or under- coverage probabilities. It is shown that uniformity can be achieved by cross-sectional bootstrap. Later, the common factors are estimated using principal component analysis (PCA), and the corresponding convergence rates are obtained. A Monte Carlo study indicates good performance in terms of mean squared error. The methodology is applied to analyze the determinants of growth rates in OECD countries.
