Topic: Contributions in panel data econometrics Title: Consistent estimation in dynamic panel data models with individual fixed effects Authors:  Daniel Becker - University of Bonn (Germany) [presenting]
Abstract: We consider the estimation of dynamic panel data models in the presence of incidental parameters for individuals. A well-known result is that the within-group estimator is biased under fixed time dimension $T$. On the other hand, it is known that the asymptotic theory of GMM estimators breaks down if the number of moment conditions divided by $T$ tends to a nonzero constant. We follow an approach using a factor analytical method. We propose estimation procedures for the model in levels and in differences. In levels we estimate the sample variance of individual effects rather than the effects themselves. We control for the initial values and correlation between the individual effects and the initial values as well as the correlation between the fixed effects and the additional exogenous variables. In differences the individual effects cancel out and we only need to control for the initial values. The estimators are consistent if $N$ tends to infinity whether or not $T$ is fixed or tends to infinity. With the proposed estimators we get rid of the well known incidental parameter problem. Our method is easy to apply, has good small sample properties and is asymptotically efficient.