The drag force experienced by coronal mass ejections (CMEs) plays an important role in determining the dynamics of CMEs. Numerous empirical studies have attempted to estimate the drag coefficient (cd) using the assumption that the observed deceleration of CMEs is primarily caused by drag alone. The observed CME motion, however, is determined by the net force—the sum of all the forces, and data contain no information to allow one to separately determine the individual forces. In the present work, we revisit the forces acting on CMEs using the erupting flux rope (EFR) model, making no assumptions on the magnitude of any force. Calculating the individual forces for four observed CME trajectories, it is shown that drag is generally not the dominant retarding force and that drag, magnetic tension, and pressure gradient can yield comparable deceleration. No force can be assumed negligible. It is further shown that the EFR equations can qualitatively replicate the observed deceleration with zero drag, contradicting the assertion that deceleration of CMEs necessarily implies the dominance of drag. A theoretical derivation of the drag coefficient cd for flux-rope CMEs is given for a range of idealized flows including the so-called “snow plow” effect. We show that cd ∼ 0–1 unless the snow-plow effect is significant or the solar wind flows are specularly reflected, in which case cd ∼ 2–3. The values of cd predicted by the EFR model for the observed trajectories are cd ≃ 1.1–1.2 for three events and cd ≃ 0.4 for one.