If a signal has a relative bandwidth of 1% I’d call it coherent. It could have a lot of noise modulation, or whatever, but that’s still coherent to me. Good commercial RF sources have RMS bandwidths in parts per trillion, but for a natural source to exhibit 1% relative bandwidth is provocative.
I accept your reasoning.
To me, having been raised in the age of lasers (I was in kindergarten when Ted Maiman announced his sensational ruby laser) I've always thought of "coherence" as being something involving parts-per-million stability.
I can see how, when one considers galactic distance and time scales, that is an unrealistic standard.
If nothing else, relative bandwidths on the order of 1% indicate a very high "Q" somewhere in the transmission chain. High "Q" in turn indicates that large amounts of electro-magnetic energy are being stored somewhere, somehow.
At least, E-M energy storage is the only cause I can think of that doesn't definitely involve intelligent design.