This paper proposes nonparametric kernel-smoothing estimation for panel data to examine the degree of heterogeneity across cross-sectional units. We first estimate the sample mean, autocovariances, and autocorrelations for each unit and then apply kernel smoothing to compute their density...

We consider instrumental variables estimation of a possibly infinite order dynamic panel autoregressive (AR) process with individual effects. The estimation is based on the sieve AR approximation with its lag order increasing with sample size. Transforming the variable to eliminate individual...