why can't the a/c stay at or above the MSA, manoeuvre to intercept the LOC a couple of miles from BN, then once on the LOC and the GS is intercepted descend with the GS?
So effectively you'd be using the glideslope intercept to substitute for a DME measurement to tell you that you were close enough to the BN to leave 3000 ft for 3200 ft? You could also do something similar with a simple DME indication, to determine that you're closer to the BN than the 1 min racetrack would take you anyway, so if it's safe to be at 3000 ft on the racetrack inbound, it must be safe to be at 3000 ft on the LLZ.
In terms of risk management, providing everyone is on top of the situation, I can't see a problem with that. The only issue is that procedures are intended to be flown as depicted, and as soon as you start taking liberties with that, you start to bet that you're capable of seeing all the failure modes of the home-made procedure. For the one time in 100 that the procedure is flown without radar, I'm not sure it's worth the stress.