This paper points out some difficulties in the econometric estimation of the original paper and suggests procedures for overcoming these.

In dynamic regression models conditional maximum likelihood (least-squares) coefficient and variance estimators are biased. From expansions of the coefficient variance and its estimator we obtain an approximation to the bias in variance es- timation and a bias corrected variance estimator, for...

This postestimation technique produces dynamic simulations of autoregressive ordinary least-squares models.