Indeed it's probably not 100%, but it's just what I observed. It is fairly close at least, nothing is really known about how HM bonuses are calculated, but what I do know is it uses class growth.
From my experience, it gives 10 "unseen" level ups for non-promoted and +20 for promoted, that's where that figure came from.
Again it's certainly not 100%, but it is a close estimate at least (only overestimating by about 1 each in comparison to your figure).
My Brenya had these stats in my last hard mode play...
HP- 42
MGC- 30*
SKL- 24
SPD- 24
LUK- 21
DEF 20*
MDF 25*
Which is actually higher then my original estimate, but I might have gotten lucky/unlucky (depending on your view). I can't remember Murdocks, but his HP was 70.
I am curious now as to how it works... if the 10 levels theory was right they would all be getting +10 in everything with the 0 RN modification... perhaps it is decided when you start a new game and you used the code after (because 0 RNs would be ideal stats for them and I have seen higher on many occasions)? Some games decide certain variables when you start a new game, if you used the cheat after that, it would be unaffected, but then again restarting the chapter changes it I think... odd indeed... it appears to be the RNG that decides it, but it isn't.
In FE4 you can do the RNG trick method to see a point of this... Use the go 1 space diagonally (2 spaces of course), have the character run, but don't wait, instead keep doing it, sometimes they will take the higher way, other times the lower way. Now in the GBA FEs you can use this to see if the RNs are low or high as you know, and you can use it to your advantage... however in FE4... it doesn't effect the RNs in the slightest..., so what causes it? Same thing with hard mode boosts, it must use an RN system separate from the main one or something, I mean there has to be a system for it, but it uses no RNs (in FE4s case, perhaps it's the same for FE6).