Inbreeding leads to reduction of the additive variance, whereas inbreeding depression reduces the performance of milk producing cows in both the nucleus and the commercial population. In this study, the cumulative additive response to 30 years of selection corrected for variance reduction due to inbreeding and inbreeding depression in the commercial cow population (denoted as expected phenotypic level or P) was evaluated in a closed (1024 cows tested per year) dairy cattle nucleus scheme, assuming a large number of gametes available per female. No dominance effects were simulated nor estimated in the nucleus. Various hierarchical and factorial designs with fewer sires than dams, an equal number of sires and dams, or even a larger number of sires than dams were compared for P. The trait considered was overall economic merit for milk production with a heritability of the unselected base population of 0·30. Sires and dams were selected on their animal model estimated additive effect for the trait considered at either 15 or 27 months of age. All full-sibs were available for selection. In the absence of inbreeding depression, a complete factorial scheme with more sires than dams resulted in the highest P. With increasing inbreeding depression, the optimal number of sires increased relatively more than the optimal number of dams. Increasing the number of sires decreased inbreeding relatively more than increasing the number of dams, and resulted in a relatively higher P. This is due to the fact that correlations between estimated additive effects of male selection candidates are higher than between those of female selection candidates.