Ordinarily no one would care if a phone has been rooted or not. It's not important. But the point here is that allowing an app that has access to your bank account to run on a rooted device is assuming that the device was intentionally rooted by it's owner. For a bank that should be assuming far too much.
Besides, even if the malware really needed to "root" the device, it could just use the same process has RootCloak[1] to hide that fact.
There's no valid security explanation; the app must designed assuming it's running on an insecure system, because there's no way of knowing if the system isn't lying.
Local app exploit that messes with the bank app only will not be detectable by the bank app.
This seem highly unlikely. In fact, I've never heard of it. And, add to that the serious hurdles I had to go through to root all the phones I ever rooted (3).
Your statements amount, to me, to overblown risk.
> A rooted phone is not a stable environment for Barclays to run their software - it could have been changed in any number of ways
This means it could be less stable, but I don't see how it would affect the possibility of a "security breach". You are talking about two different kind of issues as if they are related. Could you clarify?