In a recent paper, Mah (1994) investigated the behavior of Japanese ag
gregate import and argued that there was no empirical evidence in favo
r of the existence of a traditional imperfect-substitutes long-run imp
ort demand model using cointegration analysis. In this paper we reinve
stigate the problem in a coherent multivariate framework and point out
that the results of Mah (1994) are very much dependent on the approac
h he followed. (C) Society for Policy Modeling, 1996.