The portion of the ballast resistor voltage drop cannot have worsened by a short as long as the path to the coil primary is not shorted. If the 1.5 ohm primary resistance was measured from the ballast resistor then that is the case. This suggests that the glowing is not from the voltage drop but from the duty cycle (fraction of time the primary current is on). The ignition module has a contactless breaker, so the only way this can happen is if the ignition module is bad (or a short on the - side of the primary coil side).
1) Disconnecting the red wire in the above diagram would incapacitate the ignition module and the car should not fire.
2) Disconnecting the green wire would provide no coil path to ground, the ballast would stay cool and car should not fire.
3) Removing the ballast resistor would only have the starter relay 12V path to the coil, so the car would crank, never start, and the bendix would not disengage, which is what is reported.
4) The ballast bypass suggested is actually to short it, not remove it, and that may make the ignition module glow, so why do that.
I would try 1) and 2) just to prove Arde right, not do 4).
Then I would put a higher resistance ballast (a 12V light bulb perhaps) and put an oscilloscope on the green wire.