I'd say that it's because of the ease of designing/mapping out those generic areas that allows them to persist.
Let's say you want to make a forest. Of course, there's various types of forests, but let's narrow it down to a short list:
-generic RPG forest
-'realistic' forest [lots of leaves on the ground, etc.]
-hilly forest
-puzzlewood
Of the four here, the first is obviously the easiest to design; just throw some trees down on some grass in a vaguely mazelike pattern, maybe throwing in a river, and you're good to go. Actual forests are a bit more complex like that, obviously; you'd have to put down some leaf matter where the trees are [and those trees would definitely wouldn't be the default RPG Maker ones], various types of undergrowth to prevent players from wondering too far off, and paths to show where to actually go, but beyond that it wouldn't be a exorbitant amount of effort.
Hilly forests would be where it starts getting dicey. I can't say I've played an RPG that's done that sort of environment decently [please give me examples of this sort of stuff]. In addition to all the stuff above, you'd also have to show changes in elevation; cliffs, but also gradual rises, streams with many small waterfalls, and so on. I wouldn't wish the task of making a puzzlewood-like map of any decent size on my enemies, seeing how it's the sort of environment that actively resists implementation via 2d tiles (and I doubt parallax mapping would deal with the problem of moving through such a forest). Sure, it could be done - but would you want to spend a huge amount of time on what would end up being a relatively small area?
This problem extends to plenty of other area types as well; as areas get more 'interesting', they also get far more difficult to lay out, draw, and build. When you have an RPG's worth of areas to do that for, you don't want to spend all your time on a single portion of the game, unless that area is absurdly more important than any other area in the game (unlikely).
The problem isn't one of lack of creativity: the problem is that of implementation.