Unclebill: I hope you read my dissertation on "Stacking of Tolerances" in post #8 that took me an Hour to write?
After re reading your OP I understand what you are saying and what you are describing is similar to stacked tolerances, however what it actually is,,, is the result of a single part with multiple features made to opposite ends of tolerances. IE: one feature of the part (OD) to the minus side and one feature (ID) to the plus side. Unless this part just wore itself to these dimensions?
If you got a tighter bushing your gun would probably shoot better What I read above was entirely too loose. You have three parts that are not matched to each other, but the bushing is the primary offending item. The OD is too small for the slide and the ID is too large for the barrel. Getting an oversized or maybe just a correctly sized bushing, would probably improve it's accuracy dramatically.
In order to shoot accurately, any auto pistol must index the barrel to the slide, or more properly the sights ,,, the same way every time. On a 1911 the barrel is indexed to the slide by the fit of the locking lugs, and the barrel bushing. The link, when fit tight, will actually push up on the slide taking some of the slop out of the slide rails when in battery. This helps to insure that the rear end of the barrel returns to the same place everytime the gun is cycled.
The bushing locates the front of the barrel and if it is not tight or very close then it will flop around and not positively locate the barrel on the front end resulting in the gun shooting to a different place every shot even though the Point of Aim has not changed.
This is the simple explanation . If the two features above are correct or even close then the barrel's inherent accuracy is the only variable that is left.
Randy