This letter points out that the problems Senhadji (Applied Economics Letters, 4, 23-5, 1997) attributes to the use of the Armington aggregator in CGE models do not arise, provided that applied modellers employ the calibration procedure for CES function parameters frequently described in the literature.