Linear models with time-varying parameters: a comparison of different approaches

Estimation of linear models with time-varying parameters can be accomplished in a variety of ways, each making diferent assumptions, with varying degrees of accuracy and computational complexity. In this paper, we compare diferent gretl packages by means of simulated and real data focusing on both statistical and computational aspects. Our fndings show that all the estimators provide similar results under ideal conditions, but the practitioner’s choice could be far from obvious.