I agree with D-fendr: This is most likely software generated.
To conclude one way or another is beyond me. I must wait for someone who can ascertain for certain to show up. I take your points as to compression though, it makes a certain amount of sense.
I keep being distracted by the degree of “precision” in the gradient reproduction where edge of the original curves. I'm having trouble imagining how that edge could have been printed with toner or ink onto a hard copy with the security background and then scanned back into a digital format while managing to preserve such a remarkably smooth gradation, as if there is a level of transparency applied to it that gradually reveals less and less of the green background behind it.
Now I can produce effects very similar to that kind of apparent graduated transparency with digital illustration tools. What I can't imagine doing is printing it in such a way that it still looks like that. In printing there is either a halftone screen or dot or rosette pattern that in my experience prevents such precise variations in grayscale from being preserved on the hard copy.
Admittedly there is an X factor in that I cannot begin to guess how compression filters might handle a scan of a fine-resolution halftone screen, but it would rather surprise me if when compression is performed on a speckled background, it would wind up looking so remarkably smooth especially when so many other attributes of this have seemingly suffered a loss of clarity through compression. Yes, I know that, generally speaking, image compression can have the effect of replacing speckles with flat areas of color/shade. However I don't think I've seen such an clean illusion of transparency and such an absence of banding occur as a result of compression. Have any of you?
We all know this was a scan from an original source that didn't have a security background, but what I'm suggesting is that perhaps the original source was applied digitally to a digital image of the security background?
If so, that is an odd way to handle this, although it certainly doesn't mean that this is a forgery, nor does it cast doubt on the content of the original inset document. It just seems strange that the original would have been digitally composited with the security background instead of printed onto it, but that gradient/transparency begs me to think that may be exactly what was done.
I am no expert, but I have been around digital imaging and printing for a long time and in multiple capacities, and occasionally I'm even capable of paying attention to fine details. So, I feel almost qualified to at least ask the real experts this question. :-)
Apologies if all of this has already been kicked around elsewhere. There's just so much to read today.