Examine whether the following statements are true or false. Give an explanation.

(g) The fact that the DurbinWatson statistic is significant does not necessarily mean that there is serial correlation in the errors. One has to apply some other tests to come to this conclusion.

(h) Consider the model yt = yt 1 + xt + ut, where the errors are autoregressive. Even if the OLS method gives inconsistent estimates of the parameters, we can still use the equation for purposes of prediction if the evolution of xt during the prediction period is the same as in the estimation period.