Ordinary least squares estimation of the functional errors-in-variables model with trended data:some monte carlo evidence

This note summarizes some Monte Carlo experiments that were designed to investigate the finite sample relevance of the asymptotic normality and efficiency of OLS in the errors-in-variables model with trended data. The experiments show that the normal approximation is not very satisfactory for sample size up to T=400, but that OLS is quite efficient relative to competing estimators when there is a trend in the exogenous variables of the model.