Random House (Bertelsmann) has not cut a deal with Apple (at least nothing has been announced). It cannot be irrelevant that they are the biggest of the big 6. What does their decision tree look like?
If the "agency model" works, they can me, too at any point; they're big enough that no one can safely say no to them. "Works" would be defined as, the rest of the big 6 successfully maintain their market at the higher price. Random House might start to feel like they have to reprice -- if being cheaper doesn't gain them anything and they might not make as much money as they could. And how likely does that sound?
The "agency model" might not work, in that they don't maintain their market at the higher price. Random House, at the lower price point, might really rake it in. Not a whole lot of motivation for them to switch at that point, unless they're hurting too much from the cannibalization. Under those circumstances, Random House looks really smart to not go to the "agency model" and eventually, they buy everyone who doesn't switch back.
There is a middle ground. Random House at the lower price point might maintain their market, but not appreciably benefit from the big 5 repricing. The big 5 might _not_ maintain their market, which might buy beer or porn or Ravensburger puzzles with the money they used to spend on books. Or they might be from publishers not in the big 6, or resort to piracy. Random House is still happy about not repricing, and the rest of the big 6 don't look particularly bright. The downer here is that the industry as a whole might take a big enough hit to be felt by all book retailers, electronic and otherwise, shipping paper or pipelining e-books.
I don't see an unhappy outcome for Random House anywhere here. What did I miss?