A1436
Title: Nonstationary panel approaches to approximating nonlinear steady state functions
Authors: Peter Pedroni - Williams College (United States) [presenting]
Abstract: The aim is to propose new panel time series methods for approximating a broad class of steady state functions that are of unknown form. In particular, the asymptotic and small sample properties of a number of potential estimation approaches which involve polynomial approximation methods are investigated. These include grouped, pooled, and time-averaged cross-sectional estimators. The data-generating processes are taken to be nonstationary and heterogeneous among units of the panel. Conditions required for the approximating function to converge to the mixture average of the true function are discussed, and small sample properties are studied via Monte Carlo simulations. An empirical illustration is provided for the environmental Kuznets curve.
