It is shown that IBM cranking calculations, carried out by including a time-odd parameter in the definition of intrinsic states of IBM in the lowest order of angular frequency, give reasonably good estimates for the ground-state moment of inertia, but when extended to higher order they produce an unphysical variation of the moment of inertia with frequency. An alternative expression for the ground-state moment of inertia within the IBM framework without using time-odd parameter is briefly discussed.