My hypothesis (not being an industrial engineer) is that most inputs to panel fit are more capital & process (one-time) than operational (per-car) expenses.

If GM developed a tight-tolerance process & machinery for the Corvette, it's probably a wash (capital cost vs standardization of process) to just apply it to their other lines, albeit relaxing the specs to increase yield.

I'm not saying GM has been innovating continuously, but my experience with long-lived processes is that instantaneous improvement is far harder than continual improvement (iterating fixes), due to the specific variables and lagging feedback in play.