I don't think it works that way for RTW. In my experience, AI troops simply kill faster (or die slower) than yours at the harder battle difficulties. Unless morale helps in killing/staying alive, there's more going on than simply morale. Since my tests were one unit vs one unit, the battle ends when one unit "breaks", eliminating morale from the pre-break kill level.

A few experiments (all battles fought until 1 side routs, large battle, syrian flats during winter, no weather):

first number is my troops left at rout time, second is theirs, lower number always represents the loser

Hastati vs Hastati:

easy level:
(the AI never threw pila on easy, but my troops did)
61/10
65/32
60/13
66/15
62/14

medium:
35/6
13/21
13/54 (my general died early)
18/2
5/22

hard:
6/41
5/52
2/39
3/52
4/48

very hard:
(on very hard, my troops broke and ran much earlier than on hard, which is why there are more of them left at rout time)
17/64
11/57
18/58
11/66
24/61

As you can see, Hastati vs Hastati, all else being equal, the difficulty level does appear to dramatically affect killing ability. I'm going to try hoplites next, as their style of combat is very different.

As it stands now, you cannot expect your troops to hold their own vs equal enemy types. If that is how you like to play, that's great, but especially since Generals no longer improve unit killing/surviving ability, I don't like the fact that a unit of AI hastati might be really a unit of AI principles or heavier stats-wise on hard or very hard. I want to fight the best AI tactics, but on equal terms, unit wise. Why couldn't CA have stuck with MTW's model of leaving the stats the same, but increasing morale?