Actually, it *would* reduce costs. A significant part of the cost of new radios (particularly those manufactured in small volumes) is demonstrating compliance with the rules. That is, you have to test the radio to prove that it can't receive cell bands, even in non-standard ways (e.g. image frequencies or harmonics).
This is semantics, but I don't think that test, which is done on a single sample and not repeatedly (so it adds nothing to production cost), actually adds any compatibility lab cost at all. The receiver has to go through Part 15 testing anyway to be eligible for sale here, and the labs charge a flat fee for that, generally -- whether the sequence includes this "lockout" test or not. I used to run a compatibility/compliance lab (FCC licensed and NVLAP accredited) and we never charged extra for such tests.
It also opens up the design space. A lot of radios do not have "re-flashable" firmware because of this rule; since the FCC doesn't allow designs that can enable the prohibited bands by means of a "simple modification", which they take to mean installing/removing a jumper or diode, or loading a new software version in. As a result, you have to go to a onetime programmable part or even a mask programmed ROM.
I honestly don't see how that adds any cost, either. Whatever technology is used for releases, updates and revisions can be maintained as proprietary code not available to the public and then you can lawfully use whatever you wish to store it.