I estimated the arm power based on the measured PRG (53.6 W/W) and CARM pole (alog 65093).

Results
-------
Arm power from CARM pole = 301.0 kW
Arm power from PRG       = 337.7 kW


The first PDF details the inferences here, based on a simple coupled-cavity model supplied with some measurements we've made during commissioning.
Each measurement implies some round-trip losses, which are then used to estimate the arm power given our measured input power.  
The title of the PDF shows the assumed parameters for the model.

I also gathered some O1 parameters for comparison, taken from Hall's thesis Table 2.1 and some Kiwamu alogs.
In both cases the CARM pole tends to indicate higher round-trip losses (and therefore ~10% lower arm power) than the PRG.  
There is possibly some systematic effect in the CARM pole estimate that needs to be taken into account to make it agree with the PRG estimate (or vise versa).

Code
Plots made by /ligo/home/craig.cahillane/Git/power_budget/code/prg_from_carm_pole.py