Nonparametric Estimation of Dynamic Panel Models

This paper investigates stationary β-mixing dynamics in nonlinear panel models and develops nonparametric estimation of dynamic panel models using series approximations. We extend the standard linear dynamic panel model to a nonparametric form that maintains additive fixed effects. Convergence rates and the asymptotic distribution of the series estimator are derived, in which an asymptotic bias is present and it reduces the mean square convergence rate compared with the cross section case. Bias correction is developed using a heteroskedasticity and autocorrelation consistent (HAC) type estimator. Some extensions of this framework are also considered under exogenous variables and partial linear models. Using partial linear models, an empirical study on nonlinearity in the cross-country growth regression is presented. After bias correction, the convergence hypothesis is true only for countries in the upper income range and for OECD countries.

For more information, contact Vee Roberson.