I can kind of see where your question is coming from. But just off the top of my head, consider the following issue that can arise, as one example probably out of many similar ones:
The SI base units for pressure are $\mathrm{kg\times m^{-1}\times s^{-2}}$. Now we also know that they're directly related to the units of force in the physically meaningful sense of "pressure is force per unit area". Now, since the SI base units for force are: $\mathrm{kg\times m\times s^{-2}}$ if we adopt your suggestion, it will be perfectly legitimate, perhaps even obligatory to define the base units for pressure as: $\mathrm{kg\times m\times(m^{-2})\times s^{-2}}$, because we're dividing by area, which if I understand your suggestion will now be an independent unit.
So, the nice property of length units partially canceling will no longer be manifest in the definition of pressure, resulting in a rather longer and to my mind more cumbersome definition. While one may argue that this better preserves the physical meaning of each "piece" of the defined unit (length vs. area), I still feel that this would overall be rather confusing.
There are probably examples where this suggested convention can lead to even weirder situations, for example think of something like $\mathrm{m^2/(m^2)}$, where units that previously had completely "disappeared" from the definition, will now oddly remain present, if you see what I mean. The pressure example was just the quickest one for me to come up with.