More compression and less dynamic range is beneficial for certain environments. Noisy subways. Watching TV in a noisy downtown apartment. Basically, crappy, noisy environments. In those, compression will help you actually hear the music and speech. However, the fact that this should be done in the master is an artifact of an earlier time. Now that signal processing is small and cheap enough to be ubiquitous, music should be mastered for the best equipment, then appropriate signal processing should be done by playback.

The problem is, that there is a lot of older equipment out there that wouldn't be able to do this. So the signal gets compressed before distribution, as a compromise for the least common denominator of equipment out there. Otherwise, a big chunk of the population would think the master sounds like crap. To them, in their particular situation, it would.

EDIT: Come to think of it, the current system, where most music is more compressed, but where the people who care can still get a high dynamic range version, is a very good compromise. The problem is that the latter group's selection isn't quite filled out by the market.