Comments on "On the use of autoregressive order determination criteria in univariate white noise tests" (reply and further comments)

Comments are made on the above-named work by T.M. Pukkila and P.R. Krishnaiah (see ibid. vol.36, p.764-74, May 1988). Some new order determination criteria based on AIC, BIC, and HQ were derived theoretically. The performance was illustrated by simulations of autoregressive models. The simulation results of the old criteria are shown to be close to those of the new ones. The possible performance gain of the new criteria therefore still needs to be established experimentally. The author's reply is included, followed by further comments from S. Holm.<
>