Do you have any actual data to back up the second point?
Don't get me wrong, I understand the fill rates necessary, but i keep hearing this with no real data.
After spending the past 6-8 months working with all sorts of embedded GPU's (mainly as an HTPC hobby on embedded boards), I kinda have serious doubts that the power usage is the factor here, and would like to see real data, not assumptions.
You are assuming that higher resolution requires linear scaling of number of cores/shaders/etc, and thus, 4 times the power.
If it's displaying using 3D, it's definitely not true (it requires linear scaling of some hardware, but not all hardware, and thus, power usage is not necessarily going to be 4x).
On the 2D side, 2D cores have been so fast for the past "whatever" that it's not clear it would require any more power than normal.
In fact, the only articles i can find that did any real testing show the power usage is only increased about 30% (overall, including both panel and GPU) on a full sized laptop. Their methodology seemed very suspect, so I don't want to quote them here, just point out that there is some alternative claims.
I'm not trying to pick on you in particular, it just seems the common refrain for why the ipad mini is not retina, and i haven't seen a single shred of evidence to back it up.