autoregressive panel data model with unobserved individual effects which can
accommodate (multiple) time-invariant regressors with a large number of cross-sectional
units and a finite time periods. Our proposed methodology removes unobserved fixed effects
from the model by transforming the latter into a semiparametric additive model, however
avoids using backfitting technique. We derive the limiting results for the proposed estimators …