People need to remember that one of the reasons the "loudness wars" started in the first place was producer/label/artist A wanted his song/album to sound "louder" than producer/label/artist B. The question is, why?
A very simple answer: "louder" is almost always perceived as better. It's about standing out above the rest.
Take for example - given a set of 20 songs played in a club, all at roughly the same "loudness". Along comes one track which is "louder" than the rest. Chances are very high that more people in the club will take notice of this track. We're predispositioned to perceiving anomalies in our everyday lives, so something that is out of the ordinary (e.g. the louder track in this example) grabs our attention more than the other tracks. And at that point, the crowd would go "man, that track is really pumping".
The other issue is that the mastering engineer (who makes these kinds of calls about how "loud" or "hot" a track is before getting burnt to the master) is being paid to do something according to his client's needs. So if the producer wants the track louder, and is the one footing the bill, then there's not much the mastering engineer can do. So if the paymaster wants a loud track, that's what he will get. If mastering engineer A sticks to his guns, the producer's just going to go to another mastering house, which will mean revenue lost.
Another way to put it - if the customer wants to buy Windows NT and is dead set on this, no amount of enlightening by the consultant about the benefits of a Unix-based platform is going to change what the customer wants.
So yeah, these two factors combine and the result: the loudness wars.