OK, but what about the Audacity's conversion from stereo to "true" mono?
I guess we can not say that it does not work.
The test I did (read one of my previous comments) seems valid to me:
I compared a file having a narrower with the same file without narrower but which was converted to "true" mono with Audacity. In other words, I compared a file in "false" mono with the same file in "true" mono. To be able to compare them, I reversed the phase of one of the two files. And the phase opposition between the two files was perfect because when I mix the two files, their phase was canceled, I got a silent track.
In other words, the narrower with strength = 1 seems to perform the same transformation as the Audacity's conversion from stereo to true mono.
I understand this as a demonstration that the narrower with stregnth = 1 transforms a stereo content (the L channel is different from the R channel) into a mono content (the L channel is the same as the R channel). And therefore, the narrower can be used to check mono compatibility.
It seems then that when you say "if the Narrower worked, there should not be any change of output, if any instrument is changed in panning", it's not correct (because Audacity seems to be doing the same thing as the narrower).
If there is, where is the misunderstanding between us?
It seems like I'm passionate