Hubert - as you suggested, I ran a simple hotseat test between five France and German corps following the German occupation of Luxembourg on the first turn of the game, August 1, 1914. After the first turn, all five units stayed in place with no combat, at full strength and with supply levels from 8 - 12 (for the French unit on Verdun). I tracked their morale levels over the next ten turns.
The morale benefit for the German units was immediate: they enjoyed an immediate morale swing of 20-25% over their starting morale and remained at that level over the ten turns of my test.
The morale penalties for the French units were more ambiguous: they all dropped anywhere from 4 - 12% on the next turn (turn 2) and then rose the following turn (turn 3) to a level above their starting morale, except for the unit on Verdun which fluctuated between 5 - 11% below its starting morale of 120%.
I did a brief check of unit morale across the board and could see a similar swing in favour of the German units and against other Entente units after Luxembourg fell, but I did not document it.
These results for Germany where in range with what the manual suggests: i.e. a 20-30% morale swing in favour, but they appear to be permanent. The results for France were much less than the 20-30% swing against stated by the manual and were very temporary. I will keep track of this factor in my subsequent games and will report the results here.
But I would make the general point now that this effect seems over-powered, at least when Minors surrender. There are many more of them in the game and if attacked by a Major, in general it is not hard to force them to surrender (see the thread on Montenegro, currently at 168 posts and counting..). Would it be possible to adjust the game engine so that the Morale Effects settings can be adjusted by country; or at least to differentiate Minors from Majors in their surrender effects?
I would have to say that off hand 10 turns doesn't sound right at all. I only suggest this as I've run these kinds of tests in the past and this is the same mechanism used for all 3 SC3 games, including the WWII games and hasn't probably changed code wise in a few years.
What I would suggest is running two games side by side, where in one game you only move one CP unit into Luxembourg and in the other, don't move any units at all. If you really want to be super accurate, creating a customized campaign with all scripts removed will ensure no unforeseen events or morale events are also having an effect.
From memory it should really only be 2 turns (from a single country surrender), just about, and my guess is that by the beginning of the third CP turn, let's say around August 29th, all unit morale levels should pretty much be the same, for both the Entente and the CP, in both test games.
Actually then, since the boost happens at the end of the CP 1st turn, and everything should return to normal by the beginning of the 3rd turn, the boost in reality is only advantageous to the CP side when attacking for a single turn, which is their 2nd turn.
I only suggest this as some units will by default have high unit morale levels, 120% is possible for some units even without any boosting on both the CP and Entente side, and this is really the only way to know what is what when it comes down to the precise differences between the two test games.
If I'm wrong I'd be happy to then run this exact test myself and look into any issues that might be there.