Comments on ‘ Convergence of least squares identification algorithms applied to unstable stochastic processes ’
A counter-example is given which shows that the proof used by Fogel and Graupe (1977) to show weak consistency for the least squares method is incorrect.