I think what some people miss here is that these resistors are not just voltage dropping resistors. They are part of the Heising modulation circuit used in the radio. These resistors have an electrolytic cap across them for a reason. The resistors reduce DC power to the tube while the cap allows the AF modulation power to pass through without reduction. The overall effect is reduced carrier and increased modulation.
The D201 is not Heising modulated. The lytic across those 2 resistors keeps the resistors from wasting audio power. They only waste B+.
You are correct that the final will exhibit more drain from the xfmr by just bypassing the resistor. The plate load must be reduced (more capacitance) to where the plate dip will only give 12 watts out the pipe. This reduces plate current and most likely puts the rig where the modulation transformer ratio is optimum. It also puts the load on the power supply right back where it was at 4.1 watts out. The increased audio power needed to modulate fully will use more power but it's less than 10 watts at low duty cycle.
The compressor must be re-adjusted to obtain just under 100% negative modulation after confirming correct audio polarity. If it's only done on the bench with an audio oscillator the job isn't finished.
I've always wondered how these rigs were ever type accepted. Maybe the ones built at the very end needed to be changed to continue certification. I didn't know they changed the final toward the end. Interesting.