In another fit of sacrificing maths to the gods of procrastination I made some extensive measurements of my control arms and bushings. If you don't enjoy seeing a man wracked by indecision it is time to move on.
Bushing (Warm)
2.068” - 2.073” range = 0.005”
2.010” - 2.015” range = 0.005”
Cold
2.060” - 2.082” range = 0.022”
2.004” - 2.019” range = 0.015”
(Could the higher range be caused by ice forming on the bushing as I measured?)
Hole (Outside Diameter)
2.065” - 2.073” range = 0.008” stddev = 0.004”
1.994” - 2.015” range = 0.021” stddev = 0.007”
Warm Interference
-0.005” - 0.008”
-0.005” - 0.021”
Optimal Interference Fit
sqrt(2)*0.005 = 0.007”
0.004” - 0.007”
Inconclusions
The bushing looks fine

. I'm not thrilled with the variation in the diameter of the holes. The standard deviation isn't bad, but the extremes (particularly on the small hole) aren't great. I think these are probably a few outliers where the hole is extra stretched in one spot where something scraped on the way in or out, and its better overall.
In the worst case the fit would be 0.005" loose, but I don't think that would actually happen in practice. The other extreme is 0.021" of interference, and I think if that were to occur I wouldn't be able to fit it in there.
I'm kind of leaning towards cutting up some spacers and stuffing it all back together. I'm not sure about using loctite or not. I'll give myself the night to talk myself out of it.