I think by the time you actually finish the game, you'd be sick of the pedestrian voices. At least, that's how I've treated the GTA games I've played up until now. I can only listen to some random "I think I need some mouthwash after last night" lines so many times before I really wish I could just turn them off.
Guys the performance issue would not be relate to the storage size, but the memory, which is in really short supply on the 128mb devices (touch 2, phone 3G). There is only 20-30 megs available to an app, and there is a LOT going on in realtime already without having to manage dozens of extra sound clips for random citizen chatter. I don't think Rockstar is tying to cheat anybody here, just cut a few corners to make the game as smooth as possible on as many devices as possible. What I woul LOVE to see though is an optional download for more radio content like DJs and commercials, one of the best parts of the GTA experience.
Also just plain processing time. Uncompressed audio is fairly heavy on the CPU. If you have to spend too much time there frame rate will stink and/or you won't be able to do other stuff. But hey yeah guys, I am quite sure they are milking the franchise for money with every release. They are a business after all. Why wouldn't they? I have a hard time believing that skimping on audio is in their master plan to rule the world tho.