I guess you keep missing my point.
No, I did not. You keep making it repeatedly
If the RAV4 AWD is "more efficient", get it, only to see it gets 22/25/28 MPG from the EPA, which is less than the CX-5 AWD. In fact, there are very few AWD CUVs that get better EPA numbers than the CX-5 AWD, all with more efficient transmissions (CVT mostly). At least one of which does not get expected results on fuelly, by real owners:
Outback
Forester
2017 CR-V
Rogue
Tucson Eco (but not any other trim)
We already know that the final drive for AWD is different than the FWD, which is probably a big part of the reason there is a gap between the AWD and FWD.
Anyway, when you look at fuelly data and compare it with other vehicles, the variance of the CX-5 is equal or lower to that of competing vehicles. This suggests that the CX-5 "gap" does not materialize as much as the EPA test predicts, or that other vehicles' variance, i.e. the distribution of how many owners get crappy actual MPG, is equal or worse to that of the CX-5.