Of course, Bitcoin is also designed to charge a processing fee. This is currently optional, but it is entirely conceivable that this would change in the long run if Bitcoin ever sees adoption on a serious scale. This holds especially true given that mining awards will go down. Furthermore, services built on top of Bitcoin will have such charges as well.
I don't think economics as a science has a consensus theory that has withstood serious empirical shakedowns for explaining how high such fees end up in practice.
I'm very much behind on this since you wrote that two days ago, but I wrote my initial reply in the mindset of "what if Bitcoin becomes more pervasive" - as in: it gets increasingly used as an alternate currency. But of course your point still stands for the world today, and that other world will probably never come anyway.
I don't think economics as a science has a consensus theory that has withstood serious empirical shakedowns for explaining how high such fees end up in practice.